From 84305b11922db323ea6784ec98ca1088a3a06dd0 Mon Sep 17 00:00:00 2001
From: Luke Curley
Date: Fri, 20 Nov 2026 10:26:32 -0800
Subject: [PATCH 1/4] chore(quest): claim quest/m1/audit-docs.md
Quest-Claim: 33a197261895f2d4
From fab38dada6ce0c0ea57eb92c4f655abfbe111378 Mon Sep 17 00:00:00 2001
From: Luke Curley
Date: Fri, 20 Nov 2026 10:39:07 -0800
Subject: [PATCH 2/4] docs: run the m1 docs audit
Fix stale claims in the README, site and docs against the code, repair
the demo tape broken by the billing question and regenerate walk.gif,
document every exported item, and rewrite comments that leaned on the
TypeScript port. Plan the next run as quest/m2/audit-docs.md and add
quest/m1/missing-docs.md.
Co-Authored-By: Claude Opus 5.5
---
README.md | 4 +-
assets/demo/walk.tape | 5 ++
docs/reference.md | 40 ++++++-------
docs/releasing.md | 2 +-
docs/results.md | 9 +--
docs/semantics.md | 118 ++++++++++++++++++++-----------------
docs/t3.md | 8 +--
quest/m1/audit-docs.md | 9 ---
quest/m1/missing-docs.md | 13 ++++
quest/m2/README.md | 5 ++
quest/m2/audit-docs.md | 25 ++++++++
site/public/index.html | 7 ++-
site/public/walk.gif | Bin 717546 -> 824740 bytes
src/cli.rs | 20 ++++---
src/compaction.rs | 63 ++++++++++----------
src/context.rs | 4 +-
src/controls.rs | 17 +++---
src/discovery.rs | 34 +++++------
src/doctor.rs | 2 +-
src/exposure.rs | 43 +++++++-------
src/exposure/scan_tests.rs | 4 +-
src/fast.rs | 24 ++++----
src/idle.rs | 11 ++--
src/model.rs | 57 ++++++++++++++++++
src/prices.rs | 6 +-
src/reader.rs | 14 ++---
src/replay.rs | 8 +--
src/report.rs | 12 ++--
src/safe.rs | 4 +-
src/terminal.rs | 2 +-
src/test_support.rs | 4 +-
src/wire.rs | 1 +
tests/cli.rs | 4 +-
33 files changed, 345 insertions(+), 234 deletions(-)
delete mode 100644 quest/m1/audit-docs.md
create mode 100644 quest/m1/missing-docs.md
create mode 100644 quest/m2/README.md
create mode 100644 quest/m2/audit-docs.md
diff --git a/README.md b/README.md
index c2f6912c..05e81ea4 100644
--- a/README.md
+++ b/README.md
@@ -28,7 +28,7 @@ That is 1,190,393 Claude Code, Codex and T3 Code requests at public API
prices. The savings are what `token-max doctor --defaults` finds for a user
with no settings files on the same history: every change it would make, 8%
of the spend. Most of it is a 1h subagent cache; the rest is a 1h main-agent
-cache on usage credits, auto memory and unused skills and tools.
+cache on usage credits, auto memory and unused or costly skills and tools.
- **Subagents: a 1h prompt cache cost ~17% less than the default 5m**, ~$3,940/mo
of those savings. Every Claude Code user's subagents get 5m. The setting is
@@ -114,7 +114,7 @@ token-max doctor --help
| Tools, skills and MCP servers | What each cost to send with every request, and how often it was used |
| Instruction files | What each CLAUDE.md or AGENTS.md cost, beside the provider's size guidance |
| Fast mode | What the premium cost; recommended off |
-| Main-session effort | What each model and effort cost, beside the provider's guidance; `max` recommended down to `xhigh`, `low` worth considering raising |
+| Main-session effort | What each model and effort cost, beside the provider's guidance; `max` recommended down to `xhigh` (or `high` where a model lacks it), `low` worth considering raising |
| Auto memory | What loading, reading and writing it cost, and how often it was used; worth considering off |
| Prompt suggestions, recaps, check-ins | What each background feature cost, where it can be estimated |
| Bash timeout cap | Whether long foreground waits let the cache expire |
diff --git a/assets/demo/walk.tape b/assets/demo/walk.tape
index 50f7f20f..f1a83ee6 100644
--- a/assets/demo/walk.tape
+++ b/assets/demo/walk.tape
@@ -27,6 +27,11 @@ Show
Type "token-max doctor"
Sleep 500ms
Enter
+Sleep 2s
+# How the profile pays: usage credits, so its main agent's default cache is 5m.
+Type "a"
+Sleep 300ms
+Enter
Sleep 4s
Type "y"
Sleep 300ms
diff --git a/docs/reference.md b/docs/reference.md
index a3ddcd4e..15aead51 100644
--- a/docs/reference.md
+++ b/docs/reference.md
@@ -48,12 +48,12 @@ once, as a diff. A change whose saving was not measured, such as raising an
effort, comes after the measured ones and says so.
A change saving $5/mo or more is its own question. A smaller one joins a group
-of its kind and reason: never used, no use recorded where the window cannot
-rule one out, over `--max-cost-per-use` per use, or a redundant copy of a
-Codex catalog skill whose most-used copy stays, such as never-used skills
-or MCP servers over $1 per use. Each reason is its own group, so one yes never
+of its kind and reason, such as never-used skills or MCP servers over $1 per
+use. The reasons are never used, no use recorded where the window cannot rule
+one out, over `--max-cost-per-use` per use, and a redundant copy of a Codex
+catalog skill whose most-used copy stays. Each reason is its own group, so one yes never
covers a feature in use. A group is one question ranked by its combined saving,
-which names every member and takes `y` for all, `n` for none or `o` to go
+which names every member and takes `y` for all, `n` for none, `q` to quit or `o` to go
through them one at a time; a group of one is asked as that change. Features
whose savings were not measured, such as Codex catalog skills, are grouped the
same way after the measured questions. A group saving under $1/mo in total,
@@ -299,7 +299,7 @@ changes it, and the list ends with the source:
| --- | --- |
| Main-agent prompt cache TTL, while unset | Claude Code [requests 1h](https://code.claude.com/docs/en/prompt-caching#which-ttl-each-request-gets) for the main conversation on a Claude plan within its included usage, and 5m once usage credits or an API key pay for it. When `doctor` asked how the profile pays, the line shows that default, such as `1h (default)`, and prices against it alone. Otherwise token-max can't tell which applies, so the line shows `unset` and names both. When the replay picks a winner, setting it is the cheaper choice either way, so the line recommends `unset -> 1h` (or `5m`) and `doctor` offers to set it; it shows the replay's gap as "saves up to", saved only while the other default would apply. |
| Claude fast mode | A Claude plan [pays fast mode only from usage credits](https://code.claude.com/docs/en/fast-mode#requirements), at its API rates, even with plan usage left. |
-| Codex fast mode | A ChatGPT plan [counts Fast at 2.5x and Ultrafast at 8x](https://learn.chatgpt.com/docs/pricing) standard against included usage, and 2x and 6x on credits. The API premium priced is 2x for Fast on GPT-6 and GPT-5.6 and 6x for Ultrafast, so included usage can draw faster than it shows. |
+| Codex fast mode | A ChatGPT plan [counts Fast at 2.5x and Ultrafast at 8x](https://learn.chatgpt.com/docs/pricing) standard against included usage, and 2x and 6x on credits. The API premium priced is the [documented API rate](semantics.md#speed-controls-and-premiums), which can be lower, so included usage can draw faster than it shows. |
| Prompt suggestions | Claude Code [skips them](https://code.claude.com/docs/en/costs#background-token-usage) while a plan is near its usage limit, which the estimate, run on every eligible turn, cannot see. |
| Fable weekly use | Anthropic [caps Fable models](https://support.claude.com/en/articles/15424964-claude-fable-models-on-your-plan) at up to half the weekly usage limit on Max, Ultra and Team and seat-based Enterprise premium seats, drawing it faster than other models, runs them on usage credits on Pro and Team and seat-based Enterprise standard seats, and bills them at API rates on usage-based Enterprise. The line shows weekly API-price spend and no share of the cap, as [Fable weekly use](#fable-weekly-use) explains. |
| Codex compaction window and keepalive pings | ChatGPT credits [charge cache writes as ordinary input](https://learn.chatgpt.com/docs/pricing), while API prices from GPT-5.6 on charge 1.25x for a 30-minute write, so a replay that trades writes for reads can come out differently on credits. |
@@ -386,11 +386,12 @@ The kinds of claim are:
keepalive pings, background generation, tool output caps, and every tool,
skill and MCP server, with subagent types and context no setting turns off.
- **Official**: the provider's own guidance beside what the transcripts show:
- CLAUDE.md length and the AGENTS.md byte budget, MCP tool loading, and
- Fable's weekly use beside the plan rule. A Claude
+ CLAUDE.md length and the AGENTS.md byte budget, MCP tool loading,
+ main-session effort, and Fable's weekly use beside the plan rule. A Claude
subagent type's evidence also cites Anthropic's model guide.
- **Opinion**: the maintainer's preference, beside whatever was measured: fast
- mode off and auto memory off.
+ mode off, auto memory off, raising a `low` effort, and lowering a Codex
+ `max` effort.
The lines come in two groups. **Changes** come first: every line recommended
(red, exit 1) or worth considering (yellow, exit 0), ranked together by what
@@ -705,8 +706,7 @@ the verdict in its color, the saving green, and labels dim. `NO_COLOR`
disables color, and `--plain` also disables decorative Unicode, and either
puts the verdict word back in place of the dot. Evidence cards keep their
verdict words and only add color, and each names its claim. Redirected output
-is plain, and progress appears only on interactive stderr. SIGINT or SIGTERM ends
-with exit 2 and no success report.
+is plain, and progress appears only on interactive stderr.
Unicode decorations require a UTF-8 locale. Progress is one line that shows each
phase in turn, never an invented percent, and clears itself before the report.
@@ -718,9 +718,10 @@ as the session export below.
IDs in the report are opaque hashes. Shared histories are read once, and a
request whose profile is ambiguous counts once on the line its profiles share
-(see [Profiles](#profiles)). Independent check savings must not be summed; the
-walk's running total and the [savings against the defaults](#savings-against-the-defaults)
-are the only sums, and each says that overlapping settings may save less.
+(see [Profiles](#profiles)). Independent check savings must not be summed;
+every total the report gives, such as the walk's running total or the
+[savings against the defaults](#savings-against-the-defaults), says that
+overlapping changes may save less.
## Session export
@@ -987,7 +988,7 @@ connection that then runs the audit:
runs the first installed `token-max` whose `--version` matches, on the
non-interactive `PATH`, then in the release installers' directory (the
first of `$XDG_BIN_HOME`, `$XDG_DATA_HOME/../bin` and `~/.local/bin`),
- then in `~/.cargo/bin`. With neither, the destination downloads the
+ then in `$CARGO_HOME/bin` (default `~/.cargo/bin`). With neither, the destination downloads the
release from GitHub and checks it against its published checksum, which
needs `curl`, `tar` and `xz` there. Matching the version is not matching
the build, so a destination whose build differs from this one fails,
@@ -1020,9 +1021,8 @@ too small. Any failed
host, connection or stream before the reports are printed fails the whole run
with exit 2 and no report, and Ctrl-C then stops every destination. After
the reports, a failure during the walk, a Ctrl-C during its questions, or a
-Ctrl-C after the yes that leaves a host unwritten also exits 2, as
-[On SSH hosts](#on-ssh-hosts) describes. On a terminal, `doctor` walks through every
-host's changes and writes each on its own host, as
+Ctrl-C after the yes that leaves a host unwritten also exits 2. On a terminal,
+`doctor` walks through every host's changes and writes each on its own host, as
[On SSH hosts](#on-ssh-hosts) describes. Each host's percents
are of that host's own spend, and the total's of the hosts' spend together.
@@ -1050,8 +1050,8 @@ activity, and their metadata does not establish whether they are keepalive pings
The maintainer's cost preference is fast mode off for both harnesses. The
`fast-mode` check names the current effective local preference, its owner and
the standard-speed replacement: Claude `fastMode: false`, or Codex
-`service_tier = "default"`. Use `/fast off` for an active session. No settings
-are edited. A justified current premium preference causes exit 1 even without
+`service_tier = "default"`, which `doctor` offers to make when the profile's
+own file owns the setting. Use `/fast off` for an active session. A justified current premium preference causes exit 1 even without
historical usage or prices. Standard current settings never receive a change
because of historical premium sessions.
diff --git a/docs/releasing.md b/docs/releasing.md
index d605ecc2..53e6ece4 100644
--- a/docs/releasing.md
+++ b/docs/releasing.md
@@ -31,7 +31,7 @@ Each release carries, per target, `token-max-.tar.xz` (a
`token-max-/` directory holding the binary, README and LICENSE) or,
on Windows, a flat `token-max-.zip`, each with a `.sha256` file;
`token-max-installer.sh` and `token-max-installer.ps1`, which install into
-the first of `$XDG_BIN_HOME`, `$XDG_DATA_HOME/../bin` and `~/.local/bin`; `source.tar.gz`; and `sha256.sum` over the archives. Every
+the first of `$XDG_BIN_HOME`, `$XDG_DATA_HOME/../bin` and `~/.local/bin`; `source.tar.gz` with its `.sha256`; and `sha256.sum` over the archives. Every
asset gets a GitHub build provenance attestation, which
`gh attestation verify --repo kixel-ai/token-max` checks.
`[package.metadata.binstall]` in `Cargo.toml` points `cargo binstall` at the
diff --git a/docs/results.md b/docs/results.md
index 6caf2c5b..00575c61 100644
--- a/docs/results.md
+++ b/docs/results.md
@@ -48,10 +48,11 @@ Against the maintainer's current settings, the changes doctor still offers save
~$781/mo. Their settings already save ~$4,452/mo against the defaults, as each
report's "Compared with defaults" line gives it: the 1h caches the plan
profiles set, less the small cost of the 1h subagent cache on the usage
-credits profiles.
+credits profiles. Each figure adds up changes measured one at a time against
+a different baseline, so together they need not equal the ~$5,653/mo above.
-doctor asked how each Claude profile pays, and each answer came from the
-profile's account (`oauthAccount.billingType` in its `.claude.json`): every
+doctor asked how each Claude profile pays, and the maintainer answered from
+the profile's account (`oauthAccount.billingType` in its `.claude.json`): every
machine's `~/.claude` is on a plan, and the usage credits profiles, plus an
archived profile with no account, were answered as API. Every offered change
was declined, and no settings file changed.
@@ -71,7 +72,7 @@ The plan profiles' subagent line is now priced, so its ~$3,940/mo is in the
savings above. The previous snapshot left it out as "not enough evidence": T3
Code shares those profiles' histories, and doctor could not tell whether a T3
session ran under the profile's setting. It now reads the shared settings file
-for the controls T3 leaves alone. [How context is priced](reference.md#how-context-is-priced)
+for the controls T3 leaves alone. [T3 Code](reference.md#t3-code)
describes the rule. On the usage credits profiles, whose main agent gets 5m
unless set, 1h is the ~$518/mo in the savings; on their few subagent requests
5m was cheaper, by ~$3.63 over the window.
diff --git a/docs/semantics.md b/docs/semantics.md
index c5ad4144..3d2950d3 100644
--- a/docs/semantics.md
+++ b/docs/semantics.md
@@ -26,9 +26,9 @@ The report counts skipped lines, and `--details` the files with a deferred final
Records whose timestamp proves they precede `--since` stay outside selected
billing and usage coverage. Claude requests and compaction boundaries are retained
-separately in `scan.cacheContext` for predecessor, bracket and summary-link evidence,
+separately in `Scan.cache_context` for predecessor, bracket and summary-link evidence,
and user prompts for per-session thresholds such as a recap's three turns.
-`scan.since` records the billing selection; checks inherit it, and a comparison
+`Scan.since` records the billing selection; checks inherit it, and a comparison
cannot expand an already filtered scan to an earlier billing window.
Invalid pre-window usage becomes an unknown cache state, never a billed estimate or
silently discarded attempt; it does not fail selected billing validation. Selected
@@ -109,8 +109,8 @@ checked 2026-11-14); replays keep the API write rate, and Codex compaction and
keepalive lines say so. The
[OpenAI caching guide](https://developers.openai.com/api/docs/guides/prompt-caching),
checked 2026-11-07, documents the subset and lifetime rules. For catalog entries
-with the legacy `context_over_200k` field, supported GPT-5.4/5.5/5.6 and GPT-6 models
-use the documented 272,000-input threshold, including reads and writes. A
+with the legacy `context_over_200k` field, the GPT models `threshold` in
+`src/prices.rs` lists use the documented 272,000-input threshold, including reads and writes. A
contradictory explicit tier fails pricing. The full request is repriced when a
counterfactual crosses a tier. See [OpenAI pricing](https://developers.openai.com/api/docs/pricing).
@@ -120,7 +120,7 @@ one-hour writes derive the documented 2x ordinary-input rate. US inference uses
the documented 1.1x multiplier. Supported fast Opus rates use the documented 2x
factor across the full context window; other speed/tier combinations remain
unknown. Absent geography and the older `not_available` sentinel use the labelled
-global API-equivalent assumption, without claiming an observed region. Server web search adds its public per-search fee. Web fetch has no extra
+global API-equivalent assumption, without claiming an observed region. Server web search adds its public per-search fee, priced only on the Anthropic provider. Web fetch has no extra
fee. Code-execution charges remain unknown without runtime, tool version, and
free-allowance evidence. Client-side calls add no separate server-tool charge.
These rules and Haiku 5.5's 100,000-input threshold are from
@@ -233,8 +233,9 @@ the selected window. Sensitivity does not veto the aggregate estimate. Unsupport
remain unknown and block only their affected comparison. No attempt is silently
excluded or assigned observed zero billing.
-Current TTL resolution follows force-5m, bucket environment, bucket setting,
-enable-1h, then defaults. Shell overrides apply only to the active config root;
+Current TTL resolution follows disabled caching (no TTL), force-5m, bucket
+environment, bucket setting, enable-1h, then defaults: 5m for subagents, and for
+a main bucket on Bedrock, Vertex or Foundry. Shell overrides apply only to the active config root;
each profile's settings environment is read separately. Whether a main request
is within a Claude plan's included usage or paid by usage credits is not
inferred from credential files, so its unset default is unknown, and the line
@@ -332,7 +333,7 @@ value and the capped window. A capped user or managed global window is owned by
the model's `modelSettings` row, the only setting that changes it alone; a
capped environment or T3 window applies to every model, so it stays advisory.
Either owner then names what caps it, the variable and value or the cap itself.
-Per-model controls require v2.1.288; environment suffixes require
+Window controls require v2.1.288; environment suffixes (`k`, `m`) require
v2.1.304. Project, launch and per-agent overrides remain an explicit limitation of
the observed-file analysis. See [Claude model configuration](https://code.claude.com/docs/en/model-config#context-window-and-auto-compaction)
and [settings reference](https://code.claude.com/docs/en/settings-reference#modelsettings),
@@ -399,8 +400,8 @@ compaction on/off policy is evaluated afterward as described below.
`idle-compaction` compares on/off at current TTL/window settings and again after
applicable TTL/window changes. Each owning idle toggle gets an aggregate delta,
on cost minus off cost, across request gaps and final observation periods.
-Negative favors on, positive favors off, and exact zero retains the current
-value. Only the latter configuration decides the toggle. Independent savings
+Negative favors on, positive favors off, and exact zero is evidence for
+neither, so the line stays advisory. Only the latter configuration decides the toggle. Independent savings
are not additive. Dollars are API-price equivalents, not invoices.
Current idle controls resolve user, project, local and managed settings, with
@@ -514,8 +515,8 @@ include summary-only final periods.
## Keepalive pings
`keepalive` replays each session sequentially for no pings and every candidate
-cap. Pings fire every five sixths of the cache lifetime after a main request:
-50 minutes under a 1h TTL, below the estimated native idle-compaction trigger at
+cap. Pings fire every five sixths of the cache lifetime after a main request,
+such as 50 minutes under a 1h TTL, below the estimated native idle-compaction trigger at
54 minutes, and 25 minutes under Codex's documented 30-minute minimum. Each ping
uses the preceding main request's pricing metadata, reads its accumulated context,
and adds a prompt and reply. Prompt size is the median growth between compatible
@@ -572,7 +573,7 @@ Neither harness has a native mechanism for capped idle pings.
## Shared component measurements
-`scan.exposure` adds historical component revisions, request views, invocation
+`Scan.exposure` adds historical component revisions, request views, invocation
evidence, and client/version/window visibility. Components retain Unicode
character counts, opaque identities and fingerprints, source locations, one
owner (`skills` or `context`), and a server/namespace attribution key, or a plugin key where the record explicitly
@@ -603,7 +604,7 @@ inherited; conflicting copied histories leave exposure unknown. Current effectiv
enablement remains separately unresolved. No installed/current file size is used
to establish historical exposure.
-`report.exposure` publishes empirical calibration samples only for linked,
+`Report.exposure` publishes empirical calibration samples only for linked,
measured, single-part, complete text-only inputs with a matching request/response
model, grouped by harness, model and provider. Claude requests that record a
shape and only known text message blobs are complete. A blob holding an image,
@@ -617,7 +618,7 @@ primitive. Without matching calibration, output keeps characters and unknown
tokens/dollars. Images, opaque blocks, deferred definitions, digest-only requests,
and multi-iteration totals cannot calibrate partial text.
-`report.context` prices each piece of recurring context. Claude requests
+`Report.context` prices each piece of recurring context. Claude requests
with a recorded `api-request-shape` or full parameters qualify: tool
definitions, then the system prompt, lead the prompt in recorded order. Codex
requests qualify through their candidate `world_state` and `additional_tools`
@@ -643,7 +644,7 @@ falls in the same cache slot; otherwise it stays unpriced for that request. A
row with any unpriced piece in a request, such as an MCP server whose tool is
priced but whose instructions are not, is unpriced for that request. Each piece's calibrated
tokens are charged at the slot they fall in: cached reads, then 1h writes, then
-5m writes, then ordinary input, at the request's model, provider and context
+5m writes (for Codex, its recorded 30m or 12h writes), then ordinary input, at the request's model, provider and context
tier. A request whose calibrated pieces exceed its recorded input is left
unpriced rather than clipped. A piece's cost covers priced requests only; its
estimate scales that to every request that carried the piece and is unknown
@@ -656,8 +657,8 @@ shares conserve the aggregate's priced charge without charging it twice.
Each row scales only its own priced observations; estimates for changing
membership can have different coverage and must not be summed as a catalog
counterfactual.
-MCP tools group by server. Each server also has `details`: `mcp-tool` rows
-for individually recorded tool schemas and `mcp-discovery` rows for recorded
+MCP tools group by server. Each server also has `details`: `ContextKind::McpTool` rows
+for individually recorded tool schemas and `ContextKind::McpDiscovery` rows for recorded
discovery text. Each schema occupies its input position once, while its charge
also credits the server total. Parent coverage remains unpriced for a request
with any opaque member even when some detail rows can be priced. Tool calls
@@ -675,7 +676,7 @@ session has no request, and count on every row that shares a settings file with
them. A request no selected profile owns, such as one of a T3 instance removed from
settings, has no setting to change and prices no row; it stays in the bill.
-Tool results are `tool-output` rows, one per tool and one per MCP server. A
+Tool results are `ContextKind::ToolOutput` rows, one per tool and one per MCP server. A
result stays in the conversation until it compacts: output the previous request
held is carried, and output beyond it is new, under the bounds above. A result
that is not plain text, such as an image, leaves its row unpriced for that
@@ -721,17 +722,17 @@ transcript files. A copy cannot say which profile made a call, so it counts for
every row of its tool that shares a profile with it; skill uses count only the
model's Skill calls, since a hidden skill can still be typed, except where the
switch removes the skill entirely, where typed requests count too. A piece still sent
-today with a positive cost and no recorded use is recommended off (`verdict:
-"off"`, exit 1), whatever the amount, when the first shaped request in the
+today with a positive cost and no recorded use is recommended off (`Verdict::Off`,
+exit 1), whatever the amount, when the first shaped request in the
window of one of its profiles already carried it and none of its profiles'
transcripts could hide a call to it. An unreadable line could if its bytes name the piece (see
below), or if its transcript has a tool result whose id matches no parsed call,
a sign that a record was cut off; so could a deferred line. One every profile
first sent later, one whose profile's transcripts could hide a call, or a Codex catalog
-skill, whose uses are only partly recorded, gets `"consider"`, an advisory that exits
-0, since the window cannot rule out a use. A used piece whose `estimatedUsd`
-divided by its uses (`costPerUseUsd`) exceeds `maxCostPerUseUsd` (default $1,
-`--max-cost-per-use`) also gets `"consider"`; otherwise it is kept. The ratio
+skill, whose uses are only partly recorded, gets `Verdict::Consider`, an advisory that exits
+0, since the window cannot rule out a use. A used piece whose `estimated_usd`
+divided by its uses (`cost_per_use_usd`) exceeds `Report.max_cost_per_use_usd` (default $1,
+`--max-cost-per-use`) also gets `Verdict::Consider`; otherwise it is kept. The ratio
does not depend on the window's length. A piece with no priced request or
unresolved service-tier evidence gets no verdict. Duplicate Codex copies can
still receive a path-based recommendation without dollars. The cost is gross: removing a piece leaves later
@@ -751,7 +752,7 @@ file, the reader also hashes every tool result id and call id in the parsed
records and the whole `tool_use_id`, `call_id` and, for an output with no
`call_id`, `id` values in unreadable lines; an id cut off inside its value is not
read. Results whose id matches no parsed call, anywhere in the file and not only
-in the `--since` window, are `unmatchedResults`. Claude pairs `tool_use` blocks
+in the `--since` window, are `Unreadable.unmatched_results`. Claude pairs `tool_use` blocks
with `tool_use_id`; Codex pairs `*_call` and `*_call_output` response items by
`call_id`. The id evidence
cannot name the cut call's feature, so a file with unreadable lines and any
@@ -776,7 +777,7 @@ never retained. Different paths sharing a name remain separate rows; repeated
identical entries contribute all of their rendered characters to that path's
row, with requests counted once. Replacing aliases also changes member identity.
-`context[].skill` publishes the resolved path, requested and loaded evidence,
+`ContextCost.skill` publishes the resolved path, requested and loaded evidence,
ambiguous requests, duplicate entry count, and preferred retained path.
`calls` counts distinct observed user turns per skill, coalescing requests and
loads in that turn so they cannot inflate the cost-per-use denominator.
@@ -885,8 +886,8 @@ Calibration
estimates outside the request's measured input support stay unpriced.
No supported client/version
-currently guarantees complete typed-skill or nested-call visibility, so the
-exposure scan's `automaticUnused` stays false. Context verdicts instead rest on
+currently guarantees complete typed-skill or nested-call visibility, so
+exposure evidence alone never proves nonuse. Context verdicts instead rest on
the calls each transcript records: Claude records every tool, MCP and Skill
call, subagents' included, so a piece carried from the window's start with none
is unused. Codex catalog skills and pieces first sent mid-window stay advisory
@@ -913,7 +914,7 @@ keep once it is off, and never exits 1: notes the agent writes for itself skip
the review that checked-in `CLAUDE.md` or `AGENTS.md` gets, which encourages
slop, so the maintainer advises it off whatever it cost, as an opinion. `--max-cost-per-use` does not
apply. The check reports the total cost, the reads and writes, and the setting
-that turns it off; its `savingsUsd` is that total. Once memory is off, the line
+that turns it off; its `savings_usd` is that total. Once memory is off, the line
still shows what it cost and how often it was used earlier in the window.
Requests that recorded no context keep the memory files they were measured to
read back, and are assumed to have loaded memory as often as those that did, so
@@ -940,7 +941,7 @@ unmeasured, and the check says how many did. Today's setting decides whether
memory is on, as every other check reads today's settings. It is resolved at the
working directory of the latest main-agent request, where the user worked most
recently, including the project's `.claude/settings.json` and
-`.claude/settings.local.json`, so `offFile` names the file that owns it. Claude
+`.claude/settings.local.json`, so `off_file` names the file that owns it. Claude
Code reads the setting when a session starts, so a session begun before a change
keeps the old value until it restarts. When main-agent sessions in that
directory disagree with the setting at their latest request with a recorded
@@ -958,11 +959,11 @@ request whose reply made the call, using that model's calibrated token ratio;
writes without one stay unpriced rather than zero.
[Codex](https://developers.openai.com/codex/memories) memories are off unless
-`features.memories = true`, or both `memories.use_memories = false` (which
+`features.memories = true`, and stay off when both `memories.use_memories = false` (which
keeps them out of the prompt) and `memories.generate_memories = false` (which
stops new ones) are set. With each full context Codex 0.167.0 adds a developer message part
tagged `memories.instructions` holding `memory_summary.md`, split into several
-parts when long; earlier versions start it with `## Memory`. A full world
+parts when long, its first part starting `## Memory`. A full world
state leaves it in context. Compaction rebuilds it: the summary in the
`compacted` record's `replacement_history` takes its place, and without one it
is gone. The settings decide whether memory is on, resolved like Claude Code's
@@ -983,8 +984,8 @@ extrapolated over, or leave the cost unknown, even with memories off now, since
while memories were on. A missing summary shows nothing, since it appears only once one has
been written. Uses are
the dedicated `memories` tools (`list`, `read` and `search` as reads,
-`add_ad_hoc_note` as a write priced like a Claude write, at the 4-characters-per-token
-rule of thumb) and shell reads under `/memories/` or
+`add_ad_hoc_note` as a write priced like a Claude write, falling back to the
+4-characters-per-token rule of thumb without calibration) and shell reads under `/memories/` or
`memories_v2/`. Dedicated tool output moves to the memory row; shell output
stays on the shell's row. Extraction and consolidation run as requests Codex
records only in telemetry, so their cost is not measured: wherever memories may
@@ -1042,8 +1043,8 @@ response before the next user prompt, or the session's last response. Only
interactive terminal turns count (`entrypoint` `cli`, `querySource`
`repl_main_thread` or absent); SDK and desktop clients run suggestions only when
they ask for them, which transcripts do not record, so those turns are counted
-in the uncertainty instead. A suggestion follows a turn with at least two
-responses whose last response sent and wrote at most 10,000 uncached tokens,
+in the uncertainty instead. A suggestion follows any turn after the session's first
+response when its last response sent and wrote at most 10,000 uncached tokens,
the client's cold-cache cutoff. A recap follows a turn once the session has three
user turns and the next prompt came at least three minutes later. Each later
recap requires two new user turns after the preceding modeled recap. A final
@@ -1064,15 +1065,15 @@ live prefix refreshes it at response completion, which can keep a later recap
warm. A recap that reads a live prefix refreshes it at the recap's start. When
turning a feature off expires the prefix before the other feature or the next
recorded request reads it, the extra ordinary recap input or cache rewrite is
-subtracted from the saving as `lostWarming`. Only reads
+subtracted from the saving as `lost_warming`. Only reads
on the same model, speed and provider count, up to the prefix the recap read. Turns without
recorded usage are scaled from the priced ones, and with none priced the check
is unknown.
Uses (an accepted suggestion, a read recap) are not recorded, so nonuse is never
proven and the check never recommends a change: a feature that cost more than
-`--max-cost-per-use` over the window is worth considering off (`action:
-"advisory"`, exit 0), and one that cost less is kept, since it cannot have cost
+`--max-cost-per-use` over the window is worth considering off (`Action::Advisory`,
+exit 0), and one that cost less is kept, since it cannot have cost
more than that per use. Goal check-ins start ordinary-looking task
notification turns and Codex does not record automatic recaps, so both report an
unknown cost and stay advisory. Each line names where to turn the feature off:
@@ -1087,14 +1088,14 @@ checked 2026-11-08.
## Speed controls and premiums
-`checks` includes the `fast-mode` variant, using the same report, rendering,
+`Report.checks` includes the `CheckResult::Speed` variant, using the same report, rendering,
evidence, action and exit policy as the other checks. Its `current` and
`recommended` values describe speed after known disabling controls;
-`replacement` names the owning setting syntax. `observed.premium` is the selected
+`replacement` names the owning setting syntax. `SpeedCheck.observed.premium` is the selected
window's rate-only premium at present API rates, with known subtotal and unknown
-total kept separate. `currentEstimate.premium` applies the current effective
+total kept separate. `current_estimate.premium` applies the current effective
speed preference to those tokens. Only the latter supplies this check's
-`savingsUsd` when it recommends a change. It is a rate-only comparison, not net
+`savings_usd` when it recommends a change. It is a rate-only comparison, not net
of cache switching. A standard current preference keeps, regardless of old
premium usage. No records means unknown dollars, not zero.
@@ -1109,7 +1110,7 @@ Its local system/user files and requirements are inspected. Legacy
new `NAME.config.toml` files require an unobserved `--profile` selection. Those
scopes and project trust/launch selection stay advisory when unresolved.
-Requested controls have their own `Request.requestedSpeed` metadata and never
+Requested controls have their own `Request.requested_speed` metadata and never
populate billed `BillingPart.tier`/`speed`. In particular the native
[Codex TokenUsageRecord schema](https://github.com/openai/codex/blob/main/codex-rs/protocol/src/protocol.rs)
contains no billed tier. Native Codex historical premiums therefore remain
@@ -1120,9 +1121,10 @@ even where the general historical cost estimate labels a standard-speed
assumption. Each sampling iteration keeps its own model and billing metadata.
The [OpenAI API price table](https://developers.openai.com/api/docs/pricing),
-checked 2026-11-07, gives supported GPT-6 and GPT-5.6 fast rates at 2x standard,
-GPT-5.5 at 2.5x and GPT-5.4 at 2x below/equal 272K input. Verified ultrafast rates
-for GPT-6 Astra and GPT-6.1 Sol are 6x across their context tiers. Older or
+checked 2026-11-07, gives fast rates at 2x standard for GPT-6 and GPT-5.6 models,
+and GPT-5.5 at 2.5x and GPT-5.4 at 2x, each only at or below 272K input; ultrafast
+is 6x across context tiers. `premium_multiplier` in `src/prices.rs` lists the
+verified models. Older or
unsupported combinations remain unknown. These API rates differ from
a ChatGPT plan's multipliers, Fast at 2.5x and Ultrafast at 8x standard against
included usage and 2x and 6x on credits
@@ -1231,16 +1233,20 @@ form, and anything else, zero or negative falls back to the default (600000 ms
maximum, 120000 ms default timeout). The effective cap is the larger of the two,
so lowering only the maximum cannot override a larger default; the replacement
names both when needed. Project and local settings are not read for this check.
+The recommended cap is 4 minutes under a 5m TTL and 55 minutes under 1h, leaving
+room for the starting request's output and the reply (`safe_cap` in
+`src/bash.rs`).
The compared TTL is the main-agent TTL check's verdict when it has one, else the
current `promptCacheTtl`, else the write lifetime of the latest selected
-main-agent request; with none, the check is advisory. Claude Code 2.1.319 moves
+main-agent request; with none, or when recorded projects set the main TTL
+differently, the check is advisory. Claude Code 2.1.319 moves
commands in async subagents with a 5m cache to the background after 3 minutes
on its own, so the subagent TTL does not set the cap.
The reader keeps three facts per Claude `Bash` call in `StructuralEvent.shell`:
`background` from `run_in_background`, `polls` when the command's words include
-`sleep`, `do` and a `while`, `until` or `for` loop, and `timedOut` when the
+`sleep`, `do` and a `while`, `until` or `for` loop, and `timed_out` when the
result's `toolUseResult.timedOutAfterMs` is a number and the message holds one
tool result. The command and its output are never retained. A wait runs from
the request that started the call to its result. It is long when it reaches the
@@ -1283,10 +1289,10 @@ checked 2026-11-14, first match winning:
A user-file top-level `effortLevel` is the older form, which Opus 5.5,
Sonnet 5.5, Haiku 5.5 and Fable 5.2 ignore. Settings accept `low` to
`xhigh`; any other value is an error.
-4. The model's default: `medium` for Fable 5.2, Opus 5.5, Sonnet 5.5 and
+4. The model's default (`claude_model` in `src/effort.rs`): `medium` for Fable 5.2, Opus 5.5, Sonnet 5.5 and
Haiku 5.5, `xhigh` for Opus 4.7, `high` for the other models in the docs'
- effort table. A model outside the table takes no level; an unknown model
- leaves the line unknown.
+ effort table. Older models the table lists without effort support take no
+ level; a model it does not list leaves the line unknown.
The level then runs at the highest level the model supports at or below it
(`xhigh` runs as `high` on Opus 4.6 and Sonnet 4.6), lowered to the lowest
@@ -1342,7 +1348,8 @@ A run's model source follows Claude Code 2.1.322: the Agent call's `model`,
then the agent definition's `model:`, then `CLAUDE_CODE_SUBAGENT_MODEL`, then
the main model. `CLAUDE_CODE_SUBAGENT_MODEL_FORCE` drops the call and definition
models, so the line then names the variable as the setting to change. A user or project agent file with a built-in's name replaces it. Built-in
-Explore and Plan say `inherit`, so the variable never reaches them; Explore is
+Explore and Plan say `inherit`, so only `CLAUDE_CODE_SUBAGENT_MODEL_FORCE` lets
+the variable reach them; Explore is
capped at Opus on the first-party API when the main model ranks above it
(Fable, Mythos), unless
`CLAUDE_CODE_DISABLE_EXPLORE_INHERIT_CAP` is set, and ran on Haiku before
@@ -1386,7 +1393,8 @@ as YAML scalars without trailing comments.
The cheapest current model is, among the provider's catalog models that are not
deprecated, call tools and reply in text, the newest of each family with the
lowest output then input rate: the newest Haiku for Anthropic, the cheapest
-current GPT for OpenAI. The reprice swaps only the model on each recorded part,
+current GPT for OpenAI (`Catalog::cheapest` in `src/prices.rs`, which also needs
+known input, output and cache-read rates). The reprice swaps only the model on each recorded part,
keeping its tokens, cache categories, speed and geography, so it assumes the
same turns. A type already on that model is not repriced. A lower effort changes
the output tokens, so its cost is not measured.
diff --git a/docs/t3.md b/docs/t3.md
index 9ec05e98..a28b198c 100644
--- a/docs/t3.md
+++ b/docs/t3.md
@@ -315,11 +315,11 @@ belongs to the native profiles unless its originator is `T3 Code` or missing,
which keeps it shared. A subagent whose root session
excluded T3 excludes it too. A session whose every T3 owner is an instance no
longer in T3's settings ran under controls that no longer exist: its requests
-stay in the bill under that instance, named on a `T3 removed` report line and
-counted in `scan.t3[].removed`, and leave every current profile's checks and
+stay in the bill under that instance, counted by instance on a `T3 removed` report
+line, and leave every current profile's checks and
context costs. An
owner still in settings but under another home stays unresolved. Each request
-that stays unresolved is counted under its cause in `scan.t3[].unresolved` and
-on a `T3 unresolved` report line. A `--t3-dir`-only selection still selects configured native
+that stays unresolved is counted under its cause on a `T3 unresolved`
+report line. A `--t3-dir`-only selection still selects configured native
history directories, not a proven T3-session-only filter; retained shared records
keep a named advisory profile rather than becoming anonymous charges.
diff --git a/quest/m1/audit-docs.md b/quest/m1/audit-docs.md
deleted file mode 100644
index 6f0d0867..00000000
--- a/quest/m1/audit-docs.md
+++ /dev/null
@@ -1,9 +0,0 @@
-# [M] Audit the docs
-
-## Goal
-
-Run the `docs` audit: `quest audit docs` prints what it checks, what each finding becomes, and how to plan its next run.
-
-## Audit
-
-- docs
diff --git a/quest/m1/missing-docs.md b/quest/m1/missing-docs.md
new file mode 100644
index 00000000..6c88ca18
--- /dev/null
+++ b/quest/m1/missing-docs.md
@@ -0,0 +1,13 @@
+# [M] Gate missing docs on the library's exports
+
+## Goal
+
+Every item `src/lib.rs` exports has a one-line doc comment, and `#![warn(missing_docs)]` in `src/lib.rs` keeps new ones from landing, since CI's clippy runs with `-D warnings`.
+
+## Plan
+
+The m1 docs audit documented every exported item, but not the 673 struct fields and 165 enum variants that `pub mod` exports. 716 of them are in `src/model.rs`.
+Only `src/main.rs` and `tests/` use the library, through a few paths. The lint can't skip fields and variants.
+
+Recommended: make the modules `pub(crate)` and export only what `main.rs` and the tests use. Then document what is still exported, and add the lint.
+The alternative is documenting every field and variant. That costs about 840 lines of comments that mostly restate the field's name.
diff --git a/quest/m2/README.md b/quest/m2/README.md
new file mode 100644
index 00000000..bd7fc5c1
--- /dev/null
+++ b/quest/m2/README.md
@@ -0,0 +1,5 @@
+# After m1
+
+## Goal
+
+The work that follows m1.
diff --git a/quest/m2/audit-docs.md b/quest/m2/audit-docs.md
new file mode 100644
index 00000000..08e28319
--- /dev/null
+++ b/quest/m2/audit-docs.md
@@ -0,0 +1,25 @@
+# [M] Audit the docs
+
+## Goal
+
+Run the `docs` audit: `quest audit docs` prints what it checks, what each finding becomes, and how to plan its next run.
+
+## Plan
+
+- Baseline: the m1 run read `309b526`. Start from `git diff 309b526 -- README.md docs site src/cli.rs`, then read the pages no change touched.
+- Check first: `docs/semantics.md` against `src/`, and the README and site figures against `docs/results.md`. Those had the most stale claims in the m1 run.
+- `assets/demo/walk.tape` breaks whenever `doctor` adds a question. Run it and look at the frames, since nothing in CI runs it.
+- Scope excludes `quest/` and the vendored `.agents/` and `.claude/` skills, which quest ships.
+- `help.openai.com` answers 403 to scripted requests; check it in a browser.
+
+## Audit
+
+- docs
+
+## Required
+
+- [m1](/quest/m1/README.md) - runs once per milestone
+
+## Scheduled
+
+- 2026-11-27 - so a fast milestone does not rerun the audit at once
diff --git a/site/public/index.html b/site/public/index.html
index 87ea4620..c6452d76 100644
--- a/site/public/index.html
+++ b/site/public/index.html
@@ -49,7 +49,7 @@ On the maintainer's last 30 days, across four machines
prices. The savings are every change token-max doctor --defaults
would make for a user with no settings files, 8% of the spend: most of it the
subagent cache, the rest the main agent's cache on usage credits, auto memory and
- unused skills and tools.
+ unused or costly skills and tools.
Subagent cache, 1h instead of 5m
@@ -83,8 +83,9 @@ How it knows
cached, when it was written and whether it had expired by the next one, then
plays the same sessions again under the other setting.
Prices both. At public API prices from models.dev, its only
- download. What it can't price stays unknown, never zero. Then it offers each
- change, shows the edit and backs the file up first.
+ network request unless you name other machines with --ssh. What it
+ can't price stays unknown, never zero. Then it offers each change, shows the
+ edit and backs the file up first unless git already keeps it.
diff --git a/site/public/walk.gif b/site/public/walk.gif
index cdf2303e442a34575ad3173399dd41a8cadfa9b6..be3c3d420e22a7743c1bdf849e6f48bd280a7081 100644
GIT binary patch
literal 824740
zcmeFY`8U-6|NsA(eP8TiF!mwKkeDnr#x{foktB_sp3+E3Qq7EIY-va(Nn=UcM^uuY
zjU_vcQc2s8qEgQxMN&SVujlLgdA~lN_j#Z1_Ye4%zUEQtfL99ZcHIZt{C__9Nr;WvvaRdWAQ5(NXov3I?)YK%BH5BmL3W|D)
z3VKRP2FfbN%BqGcD*7r$I;tv0s%oaH>c(p7Mrv9{YI^!=7TW4+#_AfTtJFjrwMqAfPN7GVA%TmY2m`v0q
zYnYNXt;l+|y1F*H`qsJ>OFbn6J+ievd5ylIy@3(U&|s~h;W|TOnxT`Kk)f@z>3S1o
zBNGE#6VvskW{##7E~`~dSDVvUTRNJpb~5v@HaBxJw{W&tz20&S)5_Y-%G%x9)_aYb
z>6*1}Yu382p)uC3aa(Kaxt8X=mgQh;<6&p#W#_3+^`AxbccmfeWIh%i^gr9RZGy6==$rHJS7jlo)g>(aXB)~YZTeC!KdQfRyZ_dm{@dJBw@=mGZh6|&^5|as
z%Rk#j?)MJfe|fvT^F>GR+Xp?eo`RbGN7I96>IbD0gO5H8KAw_39v>c^eKEc?KDqSu
z_2|OXdte#RUB>vjZ4Pj8X1Q5gP$3`y0Cayqp-?~vFkAT!{M{x1av2Cjl9*yqYX(xu
zG_+e>*)AaJdzCZms~+U5ui5#syS}>X7}+I{(g=H{IUbc=7QCR<+kzubY>iTn=v7`KtHk`byXQ@w)#`!{kGeOTM;GbK{4{bpe;d2AXevdM1wStMIvd>+_4IK%hEcfL+atB)CmhZo#GJkk(3?_mH#eUP>1FY6pDf(q6?
zyBRw_4|7G%%HQ^#+WoqD@ta*#`niYq?=Mfh&nADOqIgGnGG2bEyoEER2%8q1_Qsoz5Vd&RLa`|n?*M;w4CE|%)YN<
z2#iMOk{Qqgw6ED2!_Oxpcz->anmEq7<<_=(g5Bn8be|525-I}1oOL^)94?GR@nmQF
z)n0sG7FI+ov@)n$lSQ;NvuQ(@3EKyqTle8FZS3S_|;FYfESb2CL_*R>hd1Wt>=5~I*|jSDeVgN{~EtB^g)(;~8(EV7}0P0u~vWQ**@E80c1f)Et8Gh*%adAD4a!2DJL
zV~aAmabc~~_t6A3`|vDI*XLu3J8$orf3=p+aa;4ib=k%HO^55_{RCs#u7xq}S`GtJ
zV7H!&0^4x-*sFSpKi)0n9&CHRs+$AGA3w)h_rB)T`iiEy3+tOdT)F*a@sqgA;^%bZ
zv&5gDny}#gyoXFLv8ocOt)S}|^_RrU%g)hx9e`jr^z$O*h-`avrVx~o=NepL~rSC
zQ*D-KQ41=uvCMXjQF%74zY@Q{yIp%xE}&zph|2RZGc;y}Io>sTYCYZV_OnI8f*J#3&mOOk+2Ydv8Y8=&9^bUt
z<09;Niic;ff7xtFo%eauEj_(K&9f)O1?SCTJ^MmNXHPcupSRfG(-*cldrE?>r3yUz
zBQ(C2c6--apX%w4vHyBnT2O0q!E<0|$k(!w{#x7HJp+kpU(d*}7ie9cgL}%pmQQ(K
zuz%Jwn9}_9ti0fY3;h??&-W}pJiPez9Du8%BfO-1jX4q0r;ef4E6ui_
zt0W$+b2auF$_tsRQXi;ux9c4$Oq;7F<1R8iydD*o&DBtRE_!Y0eRQ&U?mYGAMekUz
z$7Q2)wX}hYzWaM0S1it5pyMvF1YW~c8sF-eK9~GY^$yqCf4j&!dTH|ouP2v6zFi6)
zxD<4|_sO-iZl^8`|6VUVdL_zu9;ZuEf~&Js(W_-XOwV
zjq}(z`l#%CW1Y{{om=`wpEiHLDL#5NA$H@7(dE(ax0(j7ChqTh@p|$5Z3*rgN3ikb
zxW;@_x6idbr}|#Lx1YZwJ$h~Lg^jO1hRipQ3|vdO-S_HC+WcJ^?s{t1#@DlD^DR?8
z*VCT$z5d=je@}k&divzXH;be5e=ZDMKRn;}=J(?KeE=`!BD`f_%>@b4SIk%Im%$wt
zT8V|?Ok?jc^tOdI^+9pAUH=&V;6ghYU!UXQJ+63Wp@ZUEpSPudT=nk418QM?L9F*%
zjTZ}@w88qq{rzvXe=c;<@i&A5?+HE4AKgsf8^x#kCyX3^^sowVlw9zBXS(f2Z}8xa
zlehcdSseV)$Hq65c6q
z`QsrM-zY-(OffVUr8&NhRcZrM?hcDX!otQHW1kOR+ZG>{4mQ@>4SeuDxcFFvzgg$u
z^U?pz;&7es%}ZMbJ_g-gd?GHqc_r58Q^yw>=PP;P?UqTOztUd(e7!Ju`_K7-zYhQW`34{~Nf5qrzUGn)xv8m5ZBU-=urx+I
z*3@C_JCnCWeFOyR+$w`9VdZV%tt;xkJVl<-Y=dbbSDp1iyCj(Y4)f2{A<
zvKLG5X%FuV?jQVG@pEaCPH2`2eCMh(e@!trH9tBvI9KcN>jUdp^Y8`VZ9|dIN1qLT_klzIb{#P2iXkyG1(X32unPFwbu9l1
za2AW46%8ULK6Qz8U5?Zfk)$}W9!dqRWk2W9JqZ4_2Q@vh1$-w
z2FxmG;YmYXy#;2HQ7+M7`=2#7pe*kULoX@T3N-0n>bOqX)vp!AN068`Il6Pvt`#ME
zb?P&^$8D0kpj(5BBp=+X+?UsV;8r27Bws>n{c*~q@treQ%RbxALH^n^wR=H{Z-2uY&6kK@%foOda!)jLhs8OQvKZhrv|+;hdEhh
zmfu_^bYD;XR0lz;fD74@4b|B<5-hGcD!EmQq=d*p9(c`li8~GUA2B~1Q3X?>*!Qnm
zi&56V7?%0p)(VHU5*3w6G@|pTsBx_so5%i-6}G(aGkM}#ya8)FLXM~4O2P2^t^rsp
zRmA{rn}Hl0~ZLclnmq=;n@tV371B8HYq1Z!mXtg2!0ip=9i;;vy_kBz?Y5}
zx~j(49J5C&6aOnGwktW&{C7_Nc7no84CvX?w{6|xbNBax8A}0HMKXE@?vY7ClT^nB
zWNj4GSjUH=K~md2=@jw=$&H9%
z35dJDHxf-NP7s_nfpH%dqTCc!X+0VIn{6Ph?;ZDFYCP9me8l;5B4=lL8L({&{+A&4
z!HXuab?fnO`k#AmF8@B!L{1TqQ}1T(rF$L^e%6g;^I
z-(tXc+_HB5_3Bx>3Ny=z>>qXX(b1DxhzDFz%B5i0RkPtgB2-AeA2#l+RBQ1t
zQe1!Y$FU5$a-=OR&3UMmS!cK5w8P7d!gh$V>1%tT^sat~lBH)^O^ee<|5B$O<-!e6
z)nKv>P9!IBO>5W;;|G7ILs3yF+)Sydk?d{K!e&6owq%LGh|HyVKqz{vR#!cRE4By|2CZToIbGgdPzg5cX
z#zBm1tQMn$nZ(Y^Kv`KNaP%TP_^K{>d(~XBdyazx@<<7KRya!L(Gb9wGI
z_~#B)B6Hl_g){)OjRXSqaz4QT-B(%oH7yH>6Rx1h`R=5*?!G!}gf6+g++gRd$oP+3
zkpB1lkxF>7;A19p;u0w4Wv_T)ETa?MzoyI7xAo%=1fpyd6ba(GmRmPH=p6Y)P|C+R
zg!x6nw{6n}`G2xFQp8X`Lo%U_q^4G*RIhmY73V99f4s>+nm`M4L^3)AsYcfhY8rjt
zBi1XLQ}o8wj6dWd^_au%zx`}lNdwACK2{yx!IhRqB`g(h(-w&{1cziZa*VKKy=0vK
z-ZqqVx6m1GUCwnWz$T_X+7Nx>=GMZsLGRukqrXQ^Yyg4&t2ey(Yg(tL>OGL;Ax=aMj~x4ZT?*`$Fdyl595+E#5TDAM;K-YL+Q?SAum8$Tp)i
zQY->sX5n^!F5SKiWHt;22O~3%hGbM3bp5`}x&7)gL^V$x@(i$eawoS8Y~6)w%Q4Dd
z?%bEF49n~}VHul}u-8HZudq;#IeOYkR*{Tz=8}=GO<>yP%zz6&l03YYX{o|J@dXHF
zr&-ral1Ebyb1fYyGC2d5heSuUsQx4wc3&JU>hx};JUaK|>*NF1{?m#NilZOoxaDuy
zv9;dV8k%;!bUYJK!Y$cJj^J##t)A1o5AN|-)!WCB&EvPZP_>-
z+frtIgtO8Qs&%@0QME?~PC0Md!#cC)xT#ku%Vc&eL)mguZ%uBI!C)>GH5l5XSd~+5
zEVwqj&=Bcgu}b%M==?Dk*7>yoYb7XmGL%ojlPDYsfKmxQ==E-a4nu49b@DWC7r^rx
z{d`DgM}pi6eVV`-5lLzvnOA+5zXi5dWk;&U&S={frCQ}Fl=>mpPD`im8_~nPzOZW1LCJhUSJSE$(kD$;+4>Y_SR
zNJfE%{yo4sINzoPV==omte!sw2Vg>sypEqfzj*h)*Y3S>4
zIeO;i9~CEcW9WCc=pzDn
zc}I}TMP%oibL|4cFRNyHo5+oi@+=UI6ZBk<>@mI4lxkJkrfBPUI4ZMJuZM4>QWaPC
zNSBGHglX7ZEQ;~TRcB0Po=$S@E&E0HrPDZL48EZk(pCLLhj)$V)@(Ud+Qx&x$(%sH
zDwhnTF~
zTl4%Z{D!of0bj>9w`|aPowm7o{7s>Pew>&{q!aCY(3?}m)<-pUQkvGQ2VVKMLxG?@
zOL+|UJ=+|B*j!Eq3Citb9M*@rd8Lgg!jc-}52~QlXn{QeuFFKQ?;8tHklT?GrRvPx
zpr%fgp022W9%AAsRzL5nddFII+FeUuv1#4~bicG64mhYoj#t%F~lQ`;~pE&Po}3DSBvLd#X5S6xvg
zW@uNWio#s)y~|`kcbE{Jp`;d88w|-5s-hGqEPla;9^T^(F1QXQGzq7^`E41BMvi?>yWDnp|b
zYw7HoByeSGK14~+dx#$Fl0z^KS`sK!cOD_F;_sc>RGzD#x>3XN^z2)&BCBXcS7sX{
zlYrU9NcdrWVK4)!4?MbsA%k1L^YHq%rXl!Y}
zt+DS<7GR`o<`Xw(l?@ELO%!7Tez~;ytqvARKcD
zbX|Uz#6{$OQI0P0weeviZHAOxybIR2D`CN?Y|3NL0<_Z`^lt(nWz^u8lkDo^#{{AgZ0Vc
z5P3*IY5;a-Fr26nrFO=uW=nv3Vw=Vhy49DP_1JAhf_|FJKUTjopcHK|t@rHR9||!t
z4bq>}-<+#cx(cN%mEEU+W80s(Y1O|~S;u+Vy~@EF^gTqGC*iP?cpjBo#71sB@mi}+
zg=MoAmR}Vl6M|3<4tD-J-+B7aZ-;%<){PmDf1NrK7wRc>IT`byN=0AwXmEOjZB%CH
zhe~G%$jFAuMFDuS3mSaX*bi-NGeX9%G)+4X^`*-XtsYYso<&le(-M&Xk42>7UlIOC
z$NuGfALJZHDjChi#6O~7X^chFUYYIuhk7_n}
z5jJ1jJoHGvSE%4Ar1)Dnwvu
zocmFqBQmOt+-jH6du2nI*ChsyWpm=4N~@n?{2P57q?H*6<-HGUm%1__C7fP8p=_by
z&48MTUTO}*=L~=`rKKW6>;1dol(d58Qy;oP6c){r)N(pgiTzsa^Y|olEvpjZhjg~d
z3#v0atyygt2jcS~1nQZq%$`hSk^ZUco$DI-kw;dJqzQ|hTC{RUiGp3e`<0P}h=EQ{
zD$*$;4=zWPQgwkp0n)0w*S(X*GqI#!JwhJWT72P+`*PHIc93-qu!&l55ZKm7A|tn8OujSdW@~kI~0wzq+rw7AVZ&6uwXZ
znFTAew$Db5@>H#h-p1g9OGZm{V!F+#UFWtm+5951N^T^5G{+P07vI0Xwag%Y&;LkZCMXe9(Cx#}bar
z?HO&ixX_iXpet=o8HI-Hd01@nIAiDbyN}0cBP9-~9t9g;-IzsmA6an|&lS$-B95If#GSCUzNtjLT_J@^CVF*$p**m>fU+UOW`kTkuwAU1k)s@(7z~s5_MG|ahmw`A
zqNR`xSjf8>8N^O;bq~mLGf|@aavID{#-88*IBj^VxPv;uf#BEOSYSIt55&F2ImwwW
z{fsR?>|O<{7$~63#H4W-m}w~8-BMg$g|?^=N}S`&?`5n|K@b0XG5l}f(qgmo-{8{J;Qt+5qOCq&$@%Z#a_@)1vp)Y9xa_!K
zfA2rR<#%#$!LI)TmtKbs`u%Tk$=UL!D(?RQmpXPbE-iB6-{A7=!z3={{x|P`z~%0s
z04Qw7lYhWv%2Twy>5G4ZOXr|k^?$(S8STWa{{ojAclV_?fBQ7j@P}o{p}X^c{S7WB
z>)O_i&rChOyyMtCSLCbdw|(a}?MTU0lz;m4W(8c1efbTe$hgo77<&a=*2(xxT$e0^
z=uR2iIO!8NmZcsu#X78!GBqYpD1?va7@VX%Is~aC^1uqNY!J$zlFfq|pCl`wR9%7U
zO$O<*aUX-Grnf$n&GPZ+b%$l(jn>!HTM;g0D=ryYBx^girIN-fT&?ZtU+7mn9C0cf
zomG7b`|3h;NPtF7?NL~~C*yQtHuWt=y-?x|Qw){0ZlH|F1^W5GBiLCI>=Im4KYgOL
z;?QrD0&2U|8G>GaNDkHVvl&%HlTz)a=oBuG;!T;5XKOg*EIz}KILq94XG)9gf!2m{
z(?`g4LQsxAmsWY%!_AlcVS_;n+e1B&*Lwka!m#CmZnLMMS<
z--g_16Ir-!amoLE8R1o|v+^+aE4`d!y_chaUSpPZ2NIqjgx!^_$wItA*X(iGt+@A1
zIXeEL#6X@|B@v?r3WNt8>WVtbsa0YZhpsE+xec2puhgE=RqAk7ej5jIhaudw=7vo&
zwpR}XK1@hK(FDJEs_YTK&{=zl>g{mAmA;SprAR2SnpiMnSwdI?)j!oJ6P{K9ueJ14w+;k
z+Dz0cAXgl0<3|LkVbM0h!-R4n(`mz3=!T}sdu3n+CU}h)#N$o%@L>*2v?Cc7L1t;a
zPKBXj0$VZ7A9aq6)1gsGg(e)X7bSHVwrmgFPkAB-zt-$V#BQcKC9!i#^CWuxl_*Fs
zc{40Y58j}@8lyTy*5YtAM#FO;jwqH58DYiXGD4n6Av-3uDufA-(>awS44TLz-`Z9$KP@*=H^IN!eUTo(LM}e%QB@bi>
zJByXtu6=_mc^k*ZN2JU0OkBH-twy@NZw={gYy?PJjEs$?)mqa)m(58s5w_xrM>w52
zs6G9d!T>UFsRBWzFALdU_@svRNzLP@^|0&rkWEtQATYZye6Fw}wdF!%i%GvAb4`YJ
z-39ZEZV;00tnrv8VK_?Q2S+lrW+ejSU@6`AFfkv8dT-{$a@nR@3|Yg!NRL}}S-_EC
z-blLiyeL-M?MSu3cpOLo@dhNH9#<81>;q$LBWfb{Rx{k&C)}Xm`94NfA@U
zh1=G(hEpRdRpY3}+Nonm-|el+y^`k&roc{J3SVu63pc((&fWUC)p~p;-!WBJA-9f>
zW{JVZ%OyNq7?FRYnbeag(>|FsJl3Zw&;b(={-;WW3Rq+GwoaYpap*9?_wwMjktdi1qg5M(l
zG;h7)F3vEasiMtftDS?%qc60bF*yC!lAUCjO$!L;A$g80roQUC!S>Ln@KuX}@Yy!&
zSlznMaYI6=;}{1KM5E#~-1s^LO&uPo3MfTY`>1bRoC6WrMqV0GAwProrM6FI0)3E!_){C}Zx+BQ;vYK7x5OHv*f`9R?AFV(
zZ3=NMi7$TxMbWu*?f!7=nn4P9<6;GIv?#deZKBp>oVy0pABZi8#3^tA^3$Pwxbzlb
zgoapG+KP)$zXf$CXWI@c6h_ipAE5l`2BW%|HH$g!H~ll#wAABw%~c_?3)`$a>+$hS
z7p+z&5Q=VvyT2WxfDL6L)8xoqS0LJ>QyI(iEOF+$$k(gb>N$uOd47^OW6oDPrf0hg
zyQeWj-Ih775OoBfNYO_-Fu-CWzVrg_0B8H?95|QbyFd?|B%*IHe7XSaYau;E?qo?t
zKe59{97t^BXzYO*~+pG8$>o=FT%cO
z6K1&nPsQ-xv+&VnHepeM&tV5Ivsd@h)6oXd?uf)*Dt1hYm$4DgPZ3q52qh)LoCGg>
z2CxEPT@u_U8LmSH1?z$ar1(V?lKzQ~J0hi{>r+PP`1=SWC@7)LJpp)z9i&zm>popq+OCd=nkNUP{rl#f@@+HEoDRZVDp?HX}#sbP~Ou1sX^Yav9b&
z5PBmPKPJXX8zR3`F-}ylDjhE8VzyoM3Gc#o#HJH&f}<}I2C0}afY2z1yiNhcQal#`
zk+|?TQtXu+aIXhGtqJSP)}M$yd^*-NIu7&Q1OJ$g7qin0#09XN3#)TtT@J}49KN|#RV@HH=^W|cMOnOj-h)7n4c~G%)va~5a2DKkx
zNzK*5#H*cEFyu>1UW6iH;MF}#F)<#
z1dLuV$T2Y%!4}9ERWYQHx(O2w0<+*VL5svOxeuh$9SIg*mC
zjgT8`!Z9Lxk&KDv62iztRWXX(3!kQA-Kbzoww(hNHzI>=a0R1f$OXga@M77)%tiA%uxM05xXNRr?stQZOal22e;=73^;
zfDIY*QI5#|2J#i*Cg^aa2BDW~tSZ6H0~lB*?ycwodIs$E5z@!S0c-#v2dCNMCMEFI
zRMd|ffzC?!2z!uPpiB1&2!W3F9DVH$CJK?094tc!?E(npGC)lR
z+q@5HC`EjfVhgK@CR9+a5P{}EyE5SH5TXtMJAP3gIab*I-CRM6;4c$F4rH860#^Y*
zd=dUA5yUjY4HLx<^z@T^&@&=@;;Hj>z2_b7-hhUKb`lAni3kK4`B{cF4obv*V?DWp
z+)F<9QiKa*g9t*@yfDLzgSJJVMU7SNwy!cw1MZ6m{X|&vF#b5_;%7F_mkjHq69(j4
z(XF6!RQwE)@ScoearOOVxDg5BsgQF6z&im2Q{&iALWGn;cqE6AML<87AQ3~|B-oEE
zJH3NoHvr9$l+O^+du;LZaEHzxX3yM2tmnz~)@kq6Cf*!HPL%sxnmaWo-L*;5sYNhKP8}
zB~){O4Qp{P0oc}!_`BkB5lO%$DV`(F7+QC>qyg;2fd-JF5)m$3c2;Imk+-D+!2*?$
z35O(lED`<@5tL?IN5#Ub=>%Wt{icBStC651WWwUbOZ5;#8!kaiHAcwOTb}@r*f?A=
zG+l(h!3C7OHhEHUGBThl)6WjdbpYC}3vk<|ey61Pe872)qMKDSCSl3%F}KB&e!$}j
zdV%U-48T4ui!n7Zpw~f$3Pi*V3u&)h)Wsp4y*im9=Hv+afOn9C&_wiS5q$eW
zA`J{1l;9Aea2AJfT%@lGVCLiyjw0wDedXa3F*5-&00Tecerj6(?y?x4af%53g^>Yp
zb1-a#yFxBOapZMyV5fwRmr#fS42Xp>^pXT8ptr&6Z5OyGGbbXNi2rNBT$l`Hk)h=h
z+-W(m{xUwF+~`MdM03F&B0`m14=qbhoGhFm5z5$2$Z(KeGsfEk(y`R}Ik3ptooFL_
z!T?q*DMCSbOHM_@h@b#Eu3wUeNqV@)5xch^gaEL6lL0glA){i39MN(P!mJARk?4=D
z1j0q|bPL}9~uNYwG2o+sJN_co3)qN$7HxiB10`%%Q5--M-I5hRMkIfkb^dxw0aJJjt~hOUV_#V9o`#8_0b6|N!WLj;(%$`hzN@i
zg3ZN9xfuDizr@0=N>GDn*nni{X)_`DFd;~r5K1-7i^ad@V#2wxxTK01HqtoT;ynjz
zzyfQF&~nK`1c3Jsd$;8toin?rVoiX0@kb}7VI24VpIt3^Py9hyi5d$6%@L;%DUGQ&m1SP)7K
zr5~vug02zDq?yZd@dOvG_S;-d0%-yWk2xquA~#%uo0Pwuli&gb{UcJWzx0Je#VvSQ
zytM>eNF;=Czz>i95h4cTKV#Q(_1E?`#HL&N%cl@raKkrnI-L+DOf(@P7MDeE4j6Qg
zgMa^#Xd-(=z4z8Y3^`vBp(FZOdV(l=kF7l8=gb*4CxS*jF$!F;6M+3F9DWxBZRf~m
z>4c9$r2oJTMlb3;jQ+VBH^6~v`~`k_`0F%zdYbw)`p{FCjX1O%%%KzZi)QsX!^vsD
z%=!<>Ur;(kkF-PhOWeISa^yS}R%&giD!&;p4*uf;7GXI}Uj;%+Q16K-i(=T|DMz$3
z_%MEIkv#Rz0ekhMrN{>Zmw=nuc*0+x0Kk9XEYZerZzvvtAl|ve0{SFa1Qjh4!qj|-
zmJ;}!9Ij{!?mmYw6akM!IK3|opQN`dAA07z*Qn43BNHp<@}p
zw|y3E)3wE2$ICZKH%)8Z^DKk)i_=j13!WRkD+&7mC>Xh+A*1D5LpQ!ic}H+s>kaDl
zleM+s2RSCVyH8kgHm6rH(nZgw$k#+Kq$%zN;
z2;YrR4O#XtIdKeU7f8^K8ud7)z5d|$wyQFx|EHj9+wUEQZS;4p-FKRlLqY{G9YWUy
z2*u**P<>t4=A_k;tz4^wcs)AZ)%PS%?Z)uFdYLYIZ32RYtTfKYk!Z*3Q=B-TLMstl
z!*9rjee)5kY^4r03@<+H?(&E#XjoNdQJ6yI9SWV{n#0WO}*bRP&Zz
zCL(a!R+}tX^@5WB7B%0G_*pq5Z8W78lfTRZddo-CO{
zvXLL%Dn=YHS&b<~dDwl-(d5AsOHgePe4#L=OhsANPVyzIk0B#F511Im)`UBkMW-xq
zGa_k!YcnV*XUkV3w7TSkI2I}EwW)Tbz4_h-#cWvk)O$5o#v*ZiuF@GM7%=XOK&`
zN)R8A!EC2KWI#)~ev=53izMQ0J^agFuZQF)CE*%_MwI|b{7IdM&8lOWp7TDJ@Dm=ovd!tvE<_a`pn(Be=
zig`H&jF=+a$x9)+^l;IR8fQ<`+^bt6hbO
zTRCh0u&g_lp*2zj^IcE_?URAMRH^Q7^`t;Vx;|btM?x6~+%!~_9VOB4%
zVFV+n!Ezp@Ubb361X-7obbsldOlSPK2gQue*QZ8m6>W2N+*^5@z}C0bqSN=!TD$x@
z5DnePg0EXI0V_^CT>Dx00BJ5?XBkQd=Th=vL1M6KlidFKbG-8U`LhOf5Oo8y%86iB
zhDxyLV85_vi%kC*);$yYz4alw_Src~0Uh)hSp@T{^SD;PM%)0pK-0uy`qN}>u29pF
zZPi9dPek8s%>|Y7j#|53gW{=024oJ8a-}Xhv|i%fB5}2_?Et#e5t=sLh&@YffP$=j
zonowk0WK4kE9z37)@=g`D?tS
zwIMb306V=S(IJ`LAcF^j?qBU-b}9RM3cDahR^xXpZ@9+voeG}by6PoubVHuOicl88
zB{?1h7WX)E3QP(s4G)%$Uy6R032o@zw7hw!Fc~zcYeo^;cpP
zromw*NRn=9u^UCST7RLHI@$_edbep0Z5?;AkPT83(vQC{phJCaucuwjP~w}pAXMb)
z2VIkLUsBX_8^ez#pOgk!M7lT>3_0oo)mpaibV9p9SW0xUW|#U-Xp}SFnJxpJA$O&l
ziL6iFLFV}{y^Kr9UW1yRR#h2>Ixyc;Q*0ntx7DjG<3fZRFwAz%Xckww%J$}^?dE@k
z)*o|Bq1@iy(T+MiU8G_rwA~jvULTUA?#TSk)8|$v(m)=szjxvl4m9c-jAz>W7X9f)
z3IJ&Zm)n-vDQ?>*LC>T@mF`D?RtINjqM13?F-)*-X;HC{^2j!(2w(JV9C7+ZFDaVo
z5>ppxPn+4k?S2)PC48%3s^IEaa@)x0uw=!sX1+9&Z;AFF0>esIZY2S&9zqut1edEn
z+?xB|d?<6Kos+$G=>-JI(0(g?ptmy5obGYqHZ-t9O47KRkPIG?bt*IIcLgCbP$j5S
znd)4K(k7oDiz6wQI1}QT1@H?^0yE#`7Gx-;akcw9SI4%>JU7}nPDM;tu1Q*NNuE>j
z3u(2I@lYBCe64Be3D~R@AD=n~Nsz!urcxEmW*2fGdW(pYqtjZ(JBlT4xY}oe06GsA&FZo^%(`%`reuII!c#3yIkz-j
zq^BV4t
z5N8duS@iygAiR%UhEGR=YldrawiBS03x7K;_8%~yt}W7%#UXGsu$r?ZV-<_sdUc57
zaBpp}icX!gZ(&{TvZt7itjzHsbv$=UK=NRhrb1WlJ7H6z_;6M*pXRd!8bw9GY!1~)MQI{EXx=xGk7eTZA!q29Ent*TcWyJ;K90
zE{_*ygVmJ}NHbSm0tFI}X^B8apLi_YqU}QX0rrLKQ%B?30uMQVhX@hKIhJ1qiU9;e
z`Zd>b06G~FPl2qL!PC^?T}b{$DO7a?rpe)H+}hV1
z_yau<1)1en+JI|lTMf-e0+84^M0y;;Cx@%woSl+`I4BaNP!YyNij6vAIbDi~m1M<_
z+|E*jB})YRrm`Z1S*p2QsvK_V2oG+`{y^uc2{YrT7zY+$#&7d@T!1F09}-u5kaG=a
za1l)q5O+a%*Y$uj+fN-)cTjh&dUmoHk=~SD?p&sp196>eb6GH7OBNiIAZU%*79&ao
zB0or75D*GClK~d11RF>2#=-DisleG_o~OD?I=N1R4kEBJzDWeB!Eh7F`m2(&3@YMK
z4kDrs_^FPFpzuDd%Sz$!l@<_qZic_(gLriaCTE>XT2>O3YbO-wjet_8GKdtAEe#P8
znjKaOuZt7~v>*;P2@Vmn6;ipL>H;q=fX@NPHf1kAf3cQAh6gkXw$iSf&p;AsZo6_a
zKICxMs|$R{#RjZwPHJv;Hz-0Z2xId;a*Q2V+1w66yu=N%kab9ucTg&b5$7n#c>#20
zvas^ajf{OFw`d9M51RS2U~q&Gafl{xtOF2D(9F%uLrr-K>bx_@1rFs|IpbW8M1hzC
z_yi-?N^kn-^CLsG1xS$Q0xVVD#@~sA1tSjCZcZy(@V;*p%IUk(XqE
z$I!YiMzyXL!9%bDMkyXoh9q-O#Qexgjyrc%4w$Tr%;D@XPBxS2aIJ|FK^0tDfH{qG
zkCv?b1L8oOfKCVDnfz#?*Ya)&EM_Egkj!&L3SQd4w~FA5eqM|ikt&9>%)t8^0a|xY
zz>0L#wPlBAgtTQ7F4rANveS$*dk0M^u*
zCIi&WkL<8GZaJ|$g$6@V`7v?c#p)mjb-|5-!5}0!sae^KGXz%$=@PS3y7>xq4NOV1
z)@RtVd%}qnIow>>b)9}KlG3*inPvS0p(=lvsoZNv@h+}wuxLf3kh6ENw3Wp?(zi`J
za~03BAP&O<-b9a+AD0cOoFq`pMC&$`*HN64MWPW25QNyyr7Kcg=
zT++&lHq)}^0xr0QW@Tj!nw70X=9En~UBt?EY^WG4#)p`xPhuw6jqrWDP&98S=~eZBv%!3g^+wCl9~#UwIE(Dm$Orp2%t>XGz^N
z?{3O?ih1_@(^FH{vzK$8{Wa(5=+msn^;xerJ$vd`eeQ7f$fc}j?oa=h{MVQFGao#C
z_Rfz`^_7kpe)#2j_Eh>aT}$@V-skf_&3u15>)ZWj=HIdxe9ClQaEXg7_7sV3RvHdpo}LFyDW;{N%$=ex($JJO(q&(nvKzLbvK
zb(A?f<+yCmnMTTCCdxby=2WSH5N1`b)}iNA4(pE0>^@4_uExO?*>a`jk`HbT%yK_6
z8eKWY-8~j_WGwdFSlrdI_y=PNf8^Sfs5TX1=D!@{`@awr|AHM)*??CKM!e9$@N8q{
z$9P;mtSTycG5hL^vU>$q0uRD}!=M)nweUYl
z#*vpx&b?fE^=0ORms!QXbd0=QcIPE|9G;Q?vLiKX8TM6P#;g40ua+0TDyV!_c>1N^
zyS$aVm$T77{XYF=(e|$9OUsG7kIa3w{8yc0tM1izyVq+U)U3-Wkbi!)d(-mbH7_1!m+NEWGm_#*$w&@>VPdYn@Fn&IBudgM2}}^m!HY?U{EKGDKO#Pis=}-(torw)1C0)?9w`516&?>;Q%lt*IyQKqD0aKzun67yfHMjFJ?(UVK2qUZ
z=(+R^e1Nnn@Zf4WQgLX-s^ZjgE@rWl7~AvLb4M5>qc0(qU@|2rs1Umvfcm$qWnVGP
zB`bzqtCmj<6bJXm8?n;{UXz%WvvuHPh59J7D!FiuFaw?^srudfHMtW^EwqkORPQbp
zFcTAVjO89Yn?oy8Bc<}7IyjqL;aT`5{`16)>_0Dh5Z-%XBf0N(Wjm6L%3@uG$7VG}
z0>z)a^L2gsCLHiZq|a_`GoDl^4@7AFPYXaa4Y%%HQ6GG@z7Auq^XR0#LpTzvhd=2E&Dy
zzJXtb!Kj|YUxH!%vyD3{flF(AFdP^#UDS&dUlIW)Sh;%x{*oH5iG%0gZT%7jLj&Ws
zxr$|G?067>U*F_o^j!@rQNU*#;~{5bEyN35`ZM0fG(4{7su+@x#)L&A{1f5Z&y2uf%J!8vkKlEdp!vmO9flL
zwiHyM05#y0@hu3{!xcV)%x|lv?m{^5s;=>j$^0CL$JsJ;=27o~THlEEb!J79!O@AN
zTc~CBUHh3qlP{ElFE5ixWWggo=cNKqiSn3|52EJmrif|1WMuZmcxxfYqEKJd;xRN@
zOrLUv0S|CXKT-i%8jWHrFP9E^)!0kriw}K1KC6n9(p`ViKj2RHx;=?0V%6-WXD#V5
z>_0fbkv->RT|L(YNLy{#`X|l$5u>?c_lhUxQ&0hEx7`s(*Y*`J0V5gU+0kwXjIK6U5R-&I(L%V5Y>Nc5VgQ-qc$bvfh3%G
znvo}a9Z~&4c00OSE$#~%k?tqhN%_WEtbNkhG-rFQvCsGMNuhm=sW4ZK=AV|T!$%@M
z&Y}_i*s=&0Ur2o(whynT5gD|wG;aT-Ows?-I#W^*I&gzlV+qj?-aOOan;ziwc;lcL
zpPu(trigW$(
zHKx)8{x0@;vphHHK(T$9eW4xOQM(Wf9V`1tbi2g;yBtUMU02g{DoGLDy8hnvdc7kJ
zeKUGQq)N4;M(b>ocLse?`QIt@v2`uHyre(CE_y@tEXxTub7j~Q_Sx%fClI9dYDope
znXj5uN)PV5GUfPQ%$7#=?8*N*&*(&Ea+%DS0EiMr=i8tZb{mYj?qjA(e~$S$F;f}x
zxa0~Yp4`DG61e@Nz7{hOTAx_579FEb3TH&s_;WkZ&g6b>Z)ly&B(TYSqxX(#+v_}c
z)VA%TGP+Jdm@7`f`*iXc-EH;hCKThn_zIYUKuJiB>-&=OFRtF@!g!^}&6a+@VRIw-
z8|-!q>1S2G$GXj_i(-?@<0TqG*@P1PqVNjIR%g3GoZ@8SVrlUvh2KWC+k|wdP@usS
zX@oBQbgWzVn4MfYM!RH{&u%9xExwBB9wzvT?>-sHDoex!aQnif@A0{Y!Dk{?LKOE8FVHT)6!OD|NY)7D;)%(`Wozxyr3)pxd;yNrqOYeGp7rt_
z*L$KvjcJ({B`Jcb2jO+3HRA%uXN8qc=<|?7135LcENA3N_S1&7_DX40RMfiac(e@7
zU>dvyZ|w5Ry~!zbJ$_hMw;)g8`iy5t-J&I(Tsi9eR~KlPldbRW5t+S2z=9CmRup
z(cxM-y~CuSd}C4qURK8Y4Ps7oj?j9HIF8z+fmpfPeRvGFqpbqpW3pA!Q{p}Lso9xY
zw4d4#vvCB!n0yGw4agH&^{R6wKT$5VVCG7Y?5rYz%bSuUo2Z%lN!{?_``?oNyB5!l
z;6jcEFjcN1p3<}1C*D_w_`gt8cKv`*)n@E8V(zUl?k8Xg%Pw5H_4(o^z|NA_Q5Cwf
z{H#r{k}*WD46B^_O`ub`UBzR0o!KP&;{vivCNBkQ{6(7a@IaZ^og-*<@6R{4TwpnWA*Og_Y9{ltbpq
zz+&FW`gbNLQ%QwSrJ|lY`N4BK^_A~eT|1SoV%)*M_Ld*D*<`$b`rgZRA`!E^*eXK|
zT3r?R=A|Ou7v*-&uX5TmBSa`1i96JCE25X#RM>?KXfPW}FrnjK
zwFi@7x3`iM|3X=VOvIN@2e45B*_@(w1S27%gD1|^rX-eE*vD8DBw@IV{s$kTS`2>Al|NcwIc21tzL5s21J;)>pNxTw(4(34;VfM+_qmZ~x;
z&om^L!PtvhoNH(fXakJaPxY%&iM$mR|Fb
zNP%2n8_pA{h(S;7Xj=+9OpG6^E`e=Ibs$lo{XKB(4Bl0T&`nGOrf{*kgss97rK(jV
z(`+hNj{=9msLJ;b@a669>Bbw6Exk<89xr1INfkhY2A#I70<*sd=XpfF*{i=C%HDe;
zSa+t9$Y9bw16VsY6ThBpSY)?Y+)zzqF{Eb);t|B
z6M#5V9}!)lbU{LlP%ziTY8Cp&1LzlEbg%N?b-N;%$O+qJfChSp!R$@ym&puMe1u&O
z*!WuXU_6FvcS<^~uzB>bPd~pSu
zlgz_4BdP#;+(e%}d4aLh>8FRck#vjGJH4UGcI&N12$<2WmcA`4zT+s
zh-c`>^5<(NkQj!6QU=g8O>DnZ+$1F$_P9=PN_JPE&wN*Jic}f{bLECyn@$|TUkjqk`
zCR~WzXUT&w@rMf3X8D!@HsvU`3-7TJjyY-A_z%F*S%g#=Atu2mJaO5E5M&H}3Kp53~Crj>ebg3M`W;
zW0&_OUl_wN0MM0*Iz2L{>}fEtUx;7oV%k};4w}~aoF+1a^^Bbu{}YEthjWGgjy%zO
z1;&^g|IO?CBc2@1W0+i;ZjDg-A@(=)YisIhLWmQ}*kO=Sd@xeO!}=?18(|!S=h6>(
z(s}47{rwUoOs0y6x3mhm2Lm-KX9j@gwc6Dw6Vv#B5q9!jhBkvSCM6(;aZ;p61K5iV
zF=g=V!@R`fs#uXKLhU$%tHhMR1yTB0%)YB_ey^>qLN(gXsA8Q&*du9n0K~NwGI&55
z&s!w%N?<5ro545+_;k)oUxfn2;CZz`)0psn2JGJrJxNHSeVYUNzD_zujW+Ud@d!(&
zaz%TTvioA{Ug9T!;BLcD-&FQ>sL(W3giaMPbaUlzLY5A&Yc;USuCN8p;l6!DIw0&F
zJ80{kM7}2^_62i=8#D%YhBi4-2`)rvQdKz1BeOqUIMp<3z$#PB7fc_D&9ew*ngLFy
z>dPE35r)D=!f=ftkBU1$5oz`qp<@JfpdkV7F5hq#TKd_odM!af_IJp
zpV&8e|0*62hIXn@9juouVI1qv|D8eVj8;q=?5t4*^I$Ah2uSYX%!*jvHR)l*d`1`o
zxMlv$U@%~m{17_=fuyLGxqnCKpt!p0CpxsTWR#aAe_fXZ0P;oEe6)m5&zqjI4&4nqDrlM4GoRBNA-gekqZxQ
z2pq(<@`x;+=T%S>RqkAG@NmhIp9qr{P@bzowmB-lZ$iqHnG
z)*>X0A+)s|X9g2`^{-d}djRsZC5s{u7Oi96VHhI;qtjG(1n3K=0m6VPChgd|4~qNw
zOa8e}IM$_#t+N`Q?nR^-3ZQQ-4&l9CH2x~a{Kj_NBe7W#
z;(`uFgbX(3@Dd9UE;cwM@h}mJ2&v)cXDbfA3lM2}7{rh5|
zl|PTCI*y9(7uzEO?qNYF`DnE1@92_BBCh_4e|Vj0S8PuoFY!ldwMq-254?@BhG7p>abao?BlIdjP?C#ZlqH
zpoxu5b+CV{A#73^nE9)75@5$zMJ~JT_M#nZnZ5j*lHEz*uy4+2wQ5Iy6-xdW;j6cf
zpM_t+COJaFqw>-qmp%Xcx^l9^gZH06|KGm*-~cM76FfVTMra1vY$3N>a9BRqLt~hZ
zMpE$&LMyC!)hpNFa1N`VYUbIW1Qu*vPPMH_zODp&Qqe~J#_Xg`%TYG*9l{a>qVp(u
zef#&hXYwpCYUtD4)4+ihsK@0$u$*!S1S&cyHacM^wR*#2rQ
zjsN0rXkr0(8U#PTgBONxqMn3~N_Ze0io}OL+6~4xE5g{qOM3_Sj3zx^A-#{an@C-*
zEX}C3ank@m9x8fr2-Nm*X(+TDb~bm8z3afxo-h69k!#NSrs8t@h=8Ot#ilMEg{CJl
zhx+qkxKhE4iYvYdG{;XtokUX{dq}
zp7(Nne_12AaN&aKdZ?+hY9w1Mx=|
ze|jI>>nP$ypTE<0D|qD8Ti(msa!i>+K7@3>ScEQ7g>y}y5j+_1&gZ1C$^#}#4O^?Y
zcEGEKDoj)aw0*4Xi+AdFH|y6S(0-vl+P8MzrqB1ixIEGRoRMXGDlV@vNJviR(-@>A
zF~mOTO2tg)t}6~E+2GT6M=-3*lZcflQiQKeCb^hIL>Wtb&=rcTeqyDGqGAa#SlYkxJ>go>Nube}$6%z{s;p%i2q9x!u)L9jBPK3z0zb(O;ThTh`SW@@~&Lg4+)>)J`Z+UGG+N@1GR
z?vsBng8?}C>S8pEC)U7r@{UwgWbiP31o9fyV-9nN^vkPMo&PU%Y%*T;NMwq<5{{KK
zQw~5G+!G3@d8E?X2Z-OnyCW=KdR*>sYCr7L;gtd}?fnYdZ!(%pNuRvMbJYLK<`&Y|
zJWyLzhe+c~m^jM3w$~daX4*tGXReQt_b3DawRvkB3Lv`R8~~a~OfF@_aCVk7SDPT$
zTQ18n)I{-NcUCOjt2UC2XQJdM6-tPtHNJ9-%BX2qQ(!8LUDX;Bctf!`>q(zSDain)NE}!d`TOu%;x?{K!KP)Dyrg+?v+kXUY{esRvYFD4q7?(L
z-69g15M%uox?>MzpTq}0w5E%#jC-*+TshMeYNaJ*F6AFcA61OaFkAUt?rS^(AjaCd
z6$mu_fn~2Bine|@*}cnS%4QK2Z8mw@Ae?J15#CI%8m3q+rECaOAP}o%d)H~V
zQDXHyK=OER1=dwA!|ndxy`?u=i-);d{EDH)IpxkCfOglqzMkXSKYrZ%sUGl{dJTE5
zX*^2ya9Slupzcnu+5w@=`i6mfTkRs27vmS$M8=z?H6=^h9LU8Tl+UWf358!RRHpXE
z(MVGj`=bLprTT37)y-9#9`MKD|ArpP#vwq1VFaCCe;ZDJtk1_eWy@yn#)=mUcxkn})SGP8?;>Yz0yz&ZBu&d1b-4oL!YnJQ
z?14nC-@z0IQN;ZyJNrh0(1wL6}9+z{lJKjTl6A}@)uBet0kzc2kZXS>2~rNkc{FSew_sYB~>VIM{W{BVZ1
zh9K7+vUT84a06P~6%zuNktvK_)24Q&?N~!;e>vzwuGthZ=Dffg$7IR@$w?o|V}g)-
zq_JnRSqojv!*b+a-46U;-G>UVh9PbBVR8UtE#sqG6*f|FF8vFNgbwE0u6d$zGpPIV
zVh+kJ)0=S(xf0q5))LDR>~t?I$Ny{nj4iO;h#8xB47BI-kTf5YHl8Hj?U>Z%{#o5d
zBaSM9MY)OL{mGO-CT2BD2wdV^!@XO$%WoHgI?ly~S1M{hM4IShC8}K?bPd@ahQCQwXA1o2^N0
z5M#g)(X3CB7qw%?UjW|S`ubc(E`6nS1>T{x%^iR|E^-g^m+WLPR=eusYs^&MX`|_PFKhksoKQ
z1*^*Krl~S!3b^8sCABU6RuS6Aw{_cypyq3?#vF9LzXG|Na+fvriEn5y(#d1WIKexirUQ=aM!-#F$dLFo|5XbW1bgifpg3nNM
z3BE|R5xt1GdP5)Qj+dqWd7yznz3HOu(br{jAjZzArRePUHtU}VT@5`3(2s+QR;oN6
z04XeyRdQCSZ@5EG3I72v4w2-#UYe?i+&zN+Z(lw-f?Tm=-K_crJVbj_GShe3b-T%4
zt@D$X6pxu+yWRF0NEtPT7`0d0Ee(u&QFtnNUpmvNLyJ1A;Uu)m>Y~d4jE4vnib_k*
zPtqgUMOcC3zr;R1AYS#d*G5!9!5aZwiUQ6XI7MtxC5W$d#~83h|(_s-;JAjk;sR
zZheVteun&Nbc4)xtuB{_GXbvVj|*!z2-?G2Q!uy&TA;FIj+aPgAa@U4w5YYZ@KD=7Y$dls56m1W4{jK=c|U79tx;+>LpJ!_q12XfoQ~lvOvgIO
z9Ety9T=o8-sh+;SbKeB#DkWZ=>5^{I5hE$qS-nq4KUDu~
zc<;~xRWjq7Q#$jFUs{p8pb&MSFx?Qcd;Gbp9wmY?bKA(eH_)ingSZ;qB&FS0LoI=W
zjMr_%{HByP#1wezaXq~dVM_U@ZRfh|mf7weqqlg`EO{f&VPN8xU!y+Xro$kG^a}H1
z?J2qJ73ZxdS4tV>QzcULZkzbs?e?TKbg7r^k2QW%JZHuZ`y=_Ou`sT#=9GX>^wpw`
z?Y3|O{l^^YvzKSfVUPHuk-N*Gc9UbGY&i*ZzJWltZyZ0EzGs{4rFu-qD@RTt`hd*t
z$H-qI-uDSMdoF_NGtr*h`04@De!NSX#r*HpUHTe6k;~+@^4`HRT7#BSEyoS;ox?_)
zH}aRbdZUUIrw-n5`2Za7g-H{(blV0OiB~@|ov=N4yc;;{4G=ksX0;dfM`3EoTF`#3
zlhK6wp}-sg#!ugQw`DC_!Xux2XK#@bC<2i97)un4*mKw2D2ExvUbJ-}nQemF|1nC>
ze4K#MOR%UKWTa9?>)?S#pD)F0aX%qWNgDI+Hk%)O6aYGN?aPDftnDaC%5Rc_b`7W4
zx>@*a*(r|K+dg$+7tirYs$&n|alR5W3e)TO&Qn$$gC0t&e>VJ&shPiuZF2DU3gc>R
zrHYE!HPneBx?lv~rDN(=5~X}BPM#Vlb(pwiinS+;rEGPwom}Cv-OK9r!#p#cYnC~&
z4Qvk-Xo~=3?QNnKIA^U-jY_NC%tSQLaY-vUH57gmmGN4riwV
z;gAdA+;zXkY~3;K<$U({t|>Mz%;it7sE?a6kPvKmLW(KuzxknKC#KjRW6pmJyL^^=
zQ4n>0J^T2_*emM?WBu7aT9nf%(CJjn&h{CG`5a4A+>Sp(;t>9Yuo&mySbY~@k-VgL
z%ZO-Gscn&D_l6+CQ26<#`WvH>l`=+$)=MD_d)1^NqEKXjeKE9vHd#9K(W@^k(s{L?
zmg#Xrn=tPMM%g|j=0&L%)P2M46cvmxnCyAY8sFRyFNE`58FadtI}#Fgo(hV(L;lz}
z{eE~sd?Vf4>OqCWOXr83*%0ge4)yF~srD?`>mZ7;T6Q;F-7+K$bx7#n5ah^2wPT~!
zEr@j+-DF$DTlF!e0%2?f=Y4EuXT$TZte=L7n49oV_UUf$-C^|9WKz~q)MqL?!7eI@
zIhP`xyZTS|^|PV#{saK4qd-R6-Z%4M7;6547=dZg`L58akMk44XIpke{L#B2F@Gox
zg^v5LTWAl9c0m}nH2C}wTSq*>m~WGcWd832`UpGg(Jr>`c+}`7+=ye=+U}_FPnndB
z%dR_Sy^maWJ|*jGWO(9oDie^}p`?rR=zaR&$BU$th1t&|WvgpBAKgfCgDYx`q_Vmw5X-+>f&gDj)%N%sfaVVB9{mkYTkK~-ooR^UknW5{}
z*QFF^3!~!M0L@z_40Bo*DlDWluc$i5Cc(H?S!7C7{+_6U$j{mN4^}!j!Q}&^j{Vu
zwRA98v-Cz5^-?)DkuX5XyxRzQdh
zLvt+I^)TmHzI!EKvO9W9^d9tfph?W!aM-!ne$?sYh$Gi)-Oqe{cEr}}=gV$?DWeQo
zy)4e7%-pnSRI9=*zMyRI{PusoY=0m9bJoM7O%2;WoiBcbcI%e?d+YtrZ}(uggAUq<
z@SyWM*V*e`hFAWiw9AvZbm-~9dW=trDyByjs#<~X=FYqy_vKHOWzJN2I=5nEN8y_V
z6-%8@{P6)44OOguC|DsW{cUF4T1mP6uCxT{g3VlYMN1jDSzSM+Zsb;fa$dfxr}{vN
zde!IZ!zDGxdTLHh)tuqh{#H_Zp{F+M{H#6MLEi}7r4B%;QCrIqg5{McL3#&8{@OA#A3gt=
z46VIW%tR;7Uev00(IWG7&bx1aZM9Mh?w1~Lp<|YqaP5HaMy=Np1iu~5+Bu5bCvDir
z#BWCk9ZW)dY3h=z_;x0t7jCJ)fZN9;w!*nYZ*(1#5U=p?@FHjtxaL<}6Z7C6KrKeF
z?eL~$Qo=rDDVr(&{ZUdjlTZS~o`plCVVHn-E(S=Z)%RYq?BHXe|Uw&{fY!jnWcK9f^dGN7E)GE8Hn~pke
zJ>zomjQi5YU-K9lllWJGG!GNLLFUS4+N=3?5}sSUxa=L@xdE~10auiCGXnwA1BF{J
zK&(@Eocq^vrO9zrPg##+>}y+^F&aMVMI!P*+mom%o`Zyk-=}9}&)O~*pv=GL^`Np7
z;r%1YCDQ>wr*J7}dK0J4T>9^#KCbu5Qne$`@xh2)H$?0hakG`;*ag^@ku4GvL!iJ&
z4U9SjMKd{|51}f}DPAVZC`k5iwi)8Pl2Y(P`VGZWT$sW`yebE1V|fGi)&)tSX#*N`
zOW}iTCh4)4!w~L*<$_|f&RZ9fnn0A6wezB{{dYwce>K1C>S||KmneeCciY8-dz+t~D;!k)7u_|K@wyO#{qko95n`
z7r5R;y9Ii@<9l*%8~B<1XL?;`m0g%yhMQlOB6)IUMf}}QWp}^!-o2>ARF1ec!BFcw
zrw4i*TS^?~lSe_~xEIN+z&5nH?elt}$-M4xk)0^TXftmddve1QMt5a8O)>4KVEZ4u
z5wFmMLMur|$ZZgB(gd_bnCD9{oiEj>Xf3vm)8V1~q
zOG5B+RmX@&thdADNa-?zTkFV6Z=Qom5nJA9H=yu%tZ*q-xV2;7aYs)4Pw!!Y`G-RP
zCa3h@-+Gg7R@^C8xO8{qe}D~K$Qc<4-NyFlj`w)Lw=)}4cj@2XRV4L=6K*_~9sA>B
zg5@LMQkJeo7a_RqOg!x!W(yPF?1NJ>e9lL+zbJsA0ry`Gybc7-;mJy)V8x?C_0U?4`e#SqR}rZtaa&^|-e>{OFwn
z@mv4>Mu_M%wk6fNCe%b6yLUKkf9y75OXVRD3xFyzc&x~HFbBjAts25b#
zdn6=HA!k+P*E5BgCizDp`p%B@mFIYZz#5`6YtMGOxN{TDA@Mwux!t>pcwy#xV2%db
zmzWZ}G3otio!2y`;}!MkEN610+-}nOD9me
zXabv+y@IO*j}>Zawfo*eY7L{F+3&k2$w94ZpWpUH_b!`abh||b8<>EsYfUI&^Bmw^|HOo54!w>UP(T5oywS$
zi&z|D_fiZ)$`WJjikfEP&$sEco_o1o{I=Vv5n`A$BkLnBDJ+St3>@
zQ%{^d$y|QeQ6yKic(&=5ww1hzI*vY(0l&sO=EVv+sB%rf39?a~>I;$b3Q`$#UX}lh
zc+T4y+=i|cbkltF-)DXJZ@{BEYPk~E^hd1a?=?@?iFw}jMmev`>(seZ$4l35K6K&o
zq9DbEKOBE|N&c;2$(#E05^pAEN`0hO#fE-L7&F5tEb-rK0%*Zc)DU
zd-y1=*BZ|ZTeW~Br`;2I1npccqW;uQ{y_CF#soo2?;DB$I+RT5+?m3z4kz8lg)AJ;a*SyQ{
zq2`$(oeGi9Hc{i>xa3=BMBc?_cyIH;CzQRO-xPLQeuEQ{Qk@9vQ_jRK_O(@1x<(vD
zJDWgvr*wQoIp6M|uXo)>k4BrcfL&urITJyY?@9nWDgKwc@R2k08}iDwoehehrQrCoHe5=#XQx5
zymsxUTur!%Afc?Xbvrhw4Kxc_|7L~bfL03uVX3M
zTVE9|5;&xi2XNBCD#tt=nE9CzoD1AW1&_>vsT
zpiL@i$(PC$7-e(e>pWlYP~3CBwx9pIVF1g>1hHg3b1_iiIbp0yZ^fC9g9#4)>>Y?q=S
zPV4=f(#AZ7@gE^vz>ZMX*7>9q`&9g-9f1PqBHETl45k^uUyENv3ZxL7SXx{%_pVKi
z_P(cFX^$d+OabKxF)fEWg9px!9Km21JeyxigcvqZ`O^z8jE~sBroKP%bFb7ci#5?r
z(j|L7ITGWcL9yc-%ZW*CHIHQAr(!u^V^%n5QG{znGPRQkv1zl68}he{6Q~U2!aiL(
z+DTvSn3(hgv`t{K&Sx?Uih!&kfx(rr+
zJm*dMc}nQLMuEr6>l{;icG@Q@3Tk}oe0@6+&&0;^Wg}hoKfX?z#xtPqd~LhGV=9@F
zG}-9CQx-1h=?|E0LNCo!+Ml|E4lY73n%hhz`Zq1{6I~qKa9>J2q((}AxJeh<^&k|r
zx6{On@;N0x`ReM0CxoGdrnxN2KCMs}>gzXW(#!s+8HL6xE4C>WfA}7=ijXveY31HF
zSf>iI;HQ5CnmTl$7Ge@9M6rV}>d=&-(Z~E~*J37dzgF0KtnOp$+g{&{Y(XtrUlr^;
z;v`V-qgnaD;g&N@3VG-T)$*>2yA`l0`rdxvKaDG`f5cH+!wVXQSK1dU9I{GuGf4*u
zi>6yoC-(HaWoX_t)G=|nK)Ek9^n8sD*4N_u3*5alG`aL_kl4Vg(Hy`_M;-U|B#nje
zu_VbOoR74z?Dm?bY|HQ5W+mz@TkPKKgHp
zVSdgR7OxFMu<}S20sY?IAokXR7f`A^o{4HBba^gwNa(auXb128{yu@svN}JW{k1*>
zhgaPL-FX|L&RaGRI$;|YIpY=D6Tu!eE!>li?g5rtn9D^4%a0$-f^2=^vK)xa
z1Z#m3>tG*BK^;5`ktksBdUT`-O5i~k2J@2T(*#;VlaOp1PThBu(7M@2OeU@Zwj_x0
zTDa)E6lWF1Nkq`Mb6a}avmZKP&cax0ry-F*q7R?|A#5WPGc+jU@-+=+=v^dLN-MK`
zS*9iz7Nx8mI9j%Hp&wI`dI-kvI0C?Ce7!P|j6j|riXV9^_gXjFFkuGABf87tGGC1U
z)eL!4z>o49PGmUU0aU_-Sz;JhvOz;#
zVLMR&w9wxPu!?!n6e+sZObkISI>aHwNx@VNL2{Vs%-KefgM|6uHcgq6M($g-_Qm9D#re5#5Au1H*knWCXmQO3CQRpj!V~>L96kwOFj$DM->?{
z7H6s>f;%L|xLp@V6ShRj@oT2+gOJ&&Wbk5>1zmriBEk^mlkUmkt^yt5ZBJu_1}*AC
z9l~rlC`RWl0Bf1VRA864S7Z>Au!9UzU|%OKVZU})`4my>08X`3&n}0{2h7zHY7x}!K~#zkqG$m}I`%J?71c;+;srE~mM4k5T8(;ABygtdgZOqXGn*K{%_k6iJLb>PoxPg4nTKH{?u
z@!x&{b^Z}vE+=$V`<6T4#?ZJtv+P
zelH&QQnyoXa|nUn-3Q~P_*oUt$%N~5fw#l*ojyE)sL03MuRg7OL#huFa`PoRE%|k^p
z7ag0%n7^rGhG}bq04nC;j>?I01Zx9g18}$o-Td9?4OGY2`Npi;z1E**|61lRZCZkF
zBr3xkS7e>iXt4plepUa1Qbf_653SYVi^NBMkO0gCiM6F^HJU*r}IM|Ik^fY5P^}Y{YusauxI7Ly3&6){
zrY$1lJ$B+k@zg4RLbEtwERgW$d;1i4+a3*GDyF>7fktO-EtR5?_oz7E
zEp;5k%-spS9sxHrCp0q?y+8ODsJM&GSo(a_bThVFx7!b7Ei$)P!x#@uox*r_(^+h(
zR_7r_Kio`cVtO7Ah$}Us6K_Hrbl69Ylyo@k{bS5inFT*}T)J3_KkVn_hn(6c#bb^H
zlqL~sjd-kEWv;P!8;sd#E_zjS`b{8u0|&c3nIhof_wWcS&1XI;P&14;{6%b(kx{i9
zR~ol@tr0h!gPJbI)XNDTf1drk3)*1vuDyu8eO$_y;x08MR+@#WRs~lacCCrb63-#I
z5qk5`zX`C#=A!F8J!jOY={ziqKzCeEUAE~hOLucLy0E%I2
zfG)sU&t%i>$k>iOU-RA*Mn4<8^CR)mo9{%
zD%QI>lWUvXd(9{QsQf@4zbSL#1cbX;2>W2yfOZ!Lgwv&*)+!*oBLNThUw*eBpla0R
z+JE-t!jRKp_jYXJ4lT0H{gj
z<(>VQMddh}tzLzg4ec%!T?Y4f@ZrT>WDPU)
ziTkJ7mvX6>XSwv|7hQcOLZyl~5>w^c1vEzUOdw{1!$I36AFi){pA;e0|HXd%9U|<^=QzC<%)fx0zWM3ZG+xg5G
z{dq5pdckBr1(_Kil}d6jnO&9uIr
zV(9rZ(ue6ge^kI!>vq%(?6$dq4d?FHRf8^_$)5#7RYCQ+}o$23qKGc%jW#EMK-5epvwV@yQ
zu5xMv=aZ44kMvwDMyRl`NE(u(M)1t^tkn?K7-4us&sGDE;f?gABNp-CVbSD`ZxI|N
z!#BtL09rd@S5Bt6P>9i%7Wd%(x8R@S^j97kN%%8tq6@A%LQeRjvgN^jKDVYv>q6EN
zZ(jGRQ+n%GQ&RtsT~zSo54yf>rd#%Z6-L-_x|-b2??Dx2-RZMc-H(lL8*ZM;PA%vC+CV>H>cS-2N
zJNJ$WNLjqQeK6e@bcBSb1GTDl@PG}`mOu{G9@t!30TSf+-wK@zf5_Av*XQn&D~EKK
zUeNieBwX94YtMvD%Si>yXO?w94g_C^(8LOzRKV0>-XwA887^*BkDLf0Qr{>ravw_M
z2>!eV_aa)iuMmGtj*T_Tv=_ki_=w}mz>xIHb~gT^ssnjlOE*`bt_kqRImk8oh(o~p
zbHEZ8+0Hfz{#q$M3m98(kEl>#5dT`^_Nw(Zt6R6|>6{+f3juIlJqvkrzR+Q!MnrST
z(@-W>kuXgqx*X@>uK^hH3)pf#PQ2ILSB~08MkdBkNG|uDULF0n1Q?n(QKAF!G)X6~
zO?*)h(`6Pmw1V$L;Ibg_+v;^FdV!PRxicH?$wbyNkvct^Z<{q&4g2Bc_bj1qC^cbj
z;UUw~8Bc{~u{EJ32f)K-jB}r#NPd~yph0(n-%bG?3etcAKD1P0_p$ZpW{qo^jS~lM
ztNpYr>m`RhV;K7h0tLTQkPI*GJGvQG&DN48m5`MTR6qQT{`f4+tzbD9R>&uQ1BjWY
zi6Q_t8$gbngR2-b^;LjQN^qw(*IEN)$*_^k&TV1uE{B3{O3+IMivEz$vEMVwo?rb&
zcUGAq=(Ao*&_zBBLBJE$mGGp#%vXrzN|aoUjSx8o0Qf|}&&Uy53O)H3eGeK3_rC<)
z7Dz!GYv{}hVaAG-Fcms?JSv2SU?-Xycn`+)+8gj>)Uwt3Hx~8Y@=U{L=
zcg?Zi@Wg9}HE76GZ+KQ)8??Qy2WRN^_s^crdKBIe>Zf1T5A~K6+b-bzz9h;IEu|YZ
z@!7Z5%&&BadtFQqnp{}ro^q#g`N-?&THl@jPCg!a6DyzR%T8?Cx-1WuK6RpaEi6cnrWc#oCSNq$#l$Ua{ta@63^+y77zEsVVN7NUb-#PmvuA>r;95V2Vk64ws
za}rDMo=QK-3S(8Hj)Fff);s8}Do#Y8al*c6uCl~6nS_Gu%igua}sgV^*z_WIDmUGguDa0GB
z_Qo=QZ()j_eE?8|@Je_LgABUkaw3Pf!B*FLZX@PWTx*28hoxP0qr%{-;wE#rsOo01
zD%G@38fjWc6WfH9PhO&|Xol}!=6x%;F4s9TWuFme_Ur-oz5B5e%fFac2NtN9ZUu3W
z(}V2A?zAO!^w$Sh!shXV^nz&vtz~gM6C~Nf-N{Cu8-r~`fA;6Vm&B2WllQ&6t`kxj
z?@I56YY&H=i>vgSZjqB#>|So5yTUuJE04T)s#}lnXv%XaAS5fLJcJuum*<<%`Ol$H
z5$CLEF7GnZh#&4d;iI2DHK893ulsu3&E7ZJxrn2vBXME_8%KV-Vs4Vm2KA(JyzA5q
zfoSt}a>bG>nI7jjKUs|i@m3q?{kbeogT>0q!yCZ;OFp#=QF)ZCJJDKh
z=8NSu1{}5Z(LJTcb+B#UV3K@zerNcxolg>5FJe^6M6Kka&-}*-+a(IsB~^S!e~A>VtjuTt8uEE96s6qP4C#MREdAB
zTKAw&s~lk@>v7Qu%(LQk>BVuE#b$mfx|rURl*kR#526=ZF-|0&T^Q4oYPlvPR`1tO!?K*Oc8Z2gC1kKswE-3jP+`ownZk&8%9
zQ(X>Q%JXsAkW1&2rCC@x>d(y{unr@!I4{qqJ+nnx$pF)!^dkRJAi_c;q=)GuDY8U;
zzTAM|ID>O330v&N7KQ~V^Ek;1p`vF~S>@s`-fLu<;HGJV97XunxVn&I8r;vuyNqfG
z_82Vs+k*=%0)l`u=KiqyEP@PEiFi7U5aW)Ae2)0zT=7&ClC(-gJ7g6J>x&t0J8(WZ
z<$dyczvn+KR(%rsPX|*Icrddmt+0$G6By$>^5-c#(c1*pIzoLcHh_;<
zJC$d$=)J|a5}2-Ra1uCE6=v7eR^s8@V1!O3<2FLO4?1~uZuCyLV9dmutF*G#uP|&%
zth^q#A7Px57cd^j3Q%=Z!y$LSDZ0quS&3!GXeV0GS(Y-b2KJAly)$@1kDg$QqGIgZ
zZ?Iwu49xtu3`q#%>lzBS8vO?HVOF0xys5Rj@L7S>dla-Grf$b%`pV!)vM_uKZXWIM
z3}@a}mJuaHW@|Q>{4ED5<}vl*qbCxfxzPG07j&E?xAIis<-}AgJwI8LF|SjCx)&Q_
zQxx=82wQXuuE}V64O550{?=fZzCMmPm7U@o
zI%gr5ywoc2af;g-%>aBquy5hqrW>ewq|4QDg0Qm0u(9wZwWMBGA%Qz@)#g
z*xNEjsCF2PRIPOw1VTdAB>^dnJh=BOH0uRfsNblDhvlKo15}T4qnce%GSpzH%*vp2
z_g3G3o-pn=Y=~ZHL#m0dg_*u7BtbB}3850Ugg#t0m1L=07v-l?vCA`eUPotEyVn9R
zH=TxnnimD`I2DbgPe0wUiCq7+j%X-Wg2QIi*3fMNrO0}N12E|NMMyvz>uxN|bjFS=
zjyeb)F>Tp^>lgQZ4qqMylDn4G3Z7lmwSjt;B3z_Lm7L#oQfsFwM6*(YHH}zD(H>GA
zL@N0>w}@
zQbZsOOIC@9A$Y_5aq&efO<{oSY)WLkE@sG*yf64}QVRp2x#zd&;a%MOj^(vlzoQ7X2HA8tGtJow?>_g;%?*J78o*
zj?I3jO9@+x3wvI)A*=~qHyd0KAw#ZG32JL|3w?*HV-7uA{(Hh+oBLzP&8A)%)fD*k
z+>Og^s|_>%Rf*(%uH$sVZXnM>$pF`X^m)5rPMPOt(P|lbcRPA1+pdbOW1NR^VZb#D
zyt02@`cQwGJOx~$!&c;(&@}@4ZHMDJ9IkI56Ez?dc(as?sH{g9vdnigKng>!_LqP~
zCI#{VxG^%dLAa*bF@P%$T3BYK!E?xk=KhITqlPLqO
z&Rex#C(8|Hg&!aWEr)8U@X=c<
z%q+j4uK}cI3!-?8eJW8L18x}MMPUJJmFR665u69}lwRrQ0E}jx^qK-9Wz_1bG
zsMH>e3K0q6EZYG$jXz!r;0~XG5nx(e+vm%Uh5br=UZ1Tme2Jh9c8Y4YIFBsTiRNF!
z3a^E#g+&e5RxV`vnbg^^>u|O*+-f>AF|bM|-9V5}Bg}Bjkb|D^LJvxDCI>z3b}9+;setPmLdyWOTZZA{0j{K+(>Yvsvt!GR^wxeCNOxVczIPG
zaaI_s7Tc-huS<-G%>ru{(nAWYIMK8?LfqA8TR9RFrk##q3VL3&2z+=*=IH39A1;>)
z1rehNI$OX1Bu9$-i1kgg`t)*$mpdv%{Sl<&-03nzI0rDYCZwJN+Ci_09)Sm)kOI0
ziZjhJ$#d2$p8hPF6WWy97GbKDA)=Wg-MSSRfhb-LPVUn&HI~E!d1rhjk+KE2z&yUO
zgxdc^#!6SJd~u2q8};%`(~ZJy!_Sjf(;mu
z#D;kq=kMg{l%(KyN=*Ok@rfQ4>Nb}i?&9WdN5R;zXwJA^K&KZ5_E%T{byzKx=jU+6
zpH$5O+Ju6|-6gqfnO+lmgGQ)}L-umVbEE;V3#h%Ed`tDsDsoJ~196Z{9A1Hdrr!^C
zS3uz8h3rEKkO@XAewhi=&yy^RKm@lVZgwCJt}J!Wl*sq3HkFFw_+VT<_R}gUkuQpt
zBf0H}kan^Ea!NQjS05uH<}KpI5sefD%&%P>l_gO#KU+LX@VGWSOlW$2z_OyBn~5Xbr8B{
zN^%nrOFMhX5HJ8Zuq0C)OZMISFFcxsG>S{CX}_o(u;{-)`|QHMK6
zMT#1bN`o2kg^w>zTL?r?3<-v6;p9=!xMVQmL4k*S+FBcNxhoO&=qN}Zg?+eQY)Srb
z|KBt)MR@-O?9L0JSDv3`73k0m%Z}`r?sYMu38$!=_j>3Q@_X-tXAlL2)oc
zWJv~x-k*$*KMmdI5U`S9Bno+A2_bLka+s;ge|So0p)4_13q${+VScv{tIOk?efOz_
zRt(V6c*fvk$e%2jUcmxGsfJ_(p6V{9Moi_(VRZ7@_q}|&k2{uy3`niaS=UQ?1AF(U
z!1R&E^irX2{S5h|5VQM@X@SO++iobIu>w9*`^(7fTdb#GZajZ&s~xrZle!7Jk4+Bvvv-aw&&T1ai)Z;sgpePY;-;)OaVEh8>p>(wz8
zK&NtJ6=q3=^R2sKX(?+m)~%tFIFv2^Ycm60}u&rN?s+j#d|*JQor-weMEw8H)jeHLs>_;eLKW+yA
zm&2oH+g}c3Lbt}HK$xHVk1Fk%Eb&91{lHsF@|By)JBJ^&-L
zVAW?1)NK7h282!MQ;c>S;jt&V$9|4l!2Z#GKP>wpFEiTs5vU?z<%hB^UEZflgYkcf
zI|>h#dH*;QyZgla*|t7*W3Ui^fA=tE&tUtW(z|;ulM+h{y!4rO7ksBM0}*gpdx`z8
z{oz2&6db9)@3_&v=4HFd$~~*^!uv1zNX}sO_~$xHg-bSN+>gz9a0$!(gKc@1RhgFK
z{cC#U(p3ss7<3`0W)^U1NAJ!2CDW;XRDj*roFkXz1eGn7M|1>CA>%Q~fSp&IjvWh#
zkR)gjDTY|Xr*TKw*Q~V0J6v<&t1>0={H3ql>=eC{c(rdW`|rfdVVqnTor+KzntcCQ
z_~Fssf8Pjvyt5vi7x)7-WeMMuYZbSDzf<&X14tG~9
zytabgqT)0JlMZvfIwr#<{Rd1<{#p-PFQJxsxy+SMp0OOR4>jwbJ+Z@Kq$%cD&_2hV
zjt^Sm&m;}C>~wl`nzL|}^SKFyJ+toG!CBpu&5v^}f+Fh!N5(MpMfJwF+#fj?>Fzmn
zaH=i-HtrurW@~i)cjqPe8s5GlE0OyvvEgXyy~)V^jjjjgFZ}vmKNN2r*XaC%H+t)d
z<sSnnRAE1x?9&5XE$`71+aXlAKa?NJVj8`rcD|o_?l@mSXPFj_>xWY<8->Xp7
z@}>Nv;8-ry0M}ci2&$75YS&O;JgrV7lZ*Y0Kh;m
z|N9^RAD@z6AVeDmOrf(I1!%EL0u}N=A%@^+Bh|d@D!^JgmM~`I-9<=r(lc(Hbziah
zQo(N)Wo_hsy2;VX_KK7RVQc(7ST&Ysg79eU-IE+d%i*0*%lvmlj-fUZ8^6mjC@JoIXSf#|iafnx@0g
zt-FB&7x&a|*uU<&v4f~)^6}@_1G#>M4}Uw1AG@!^zG+$2s^X%Y6%U=4o{YP>Wwpsy=ckJ0eN8UAza7I?rwf}HuuG?PgrvE3Attao
z9a}MmUgYm93o9*k>ZyVTfj4(a==|Ab^$4u^IP$-1=D+Xk|N9vK2O6GC{&RTJLiKGP
zp|LE?!{`~VVv$3|1!U`bUOO|K=G!#MyPsMl+o$Kiz=b6C9J{rj@oEX#_w3}Yy^KGn
zI+dR+i){b?eEaOg$T4H0S8V0vtr1HTnRD##{>bxhj2qz#D{0cC4I@YAuMz8Wtq?_p
z-suUF%lX|8bKO#>N1D>Q5hx;Ik#Gf?d9D+EZO$t6d7*
zC#ul+@OG18+P9flp~0*&4=F&k-MiR_Bfbs9xMqKQXtDOSPFF=6^VpUpwqwD?2OWaM
z=DNH09G^c!@azA`OhOb{eK->I%5i5)^!?CI`+IBL>paoX&j0Y>R#GwrVs!7Re=h-X8X(*I4LvLqlAK*Qajg88qyP{=!>x^P_8uc+&W
z>S94{L|gsMV&gd*pY^vj^p!#5o_XZi#>gVaMJ=nc&i*rSh`IK5ZREM8p~I~0@17NI
z`{!<@e;&?a!TILlhKT)+t5i+egdop3@%GZ@yT4(c&=#GmYo-VOv_YQSj`uLJL
zVN$FxYoD3u-yd9SC7VrqrD|9h`#1Z_*Y~5#22LNn*75!0z9WCMCpqR`xzH}twG${b
zZe8ls_hP6tS3t5{DihL-I}MyOZ~SrKb|lB
zV{E<8^QT@vV(z!*$qHA>^2zc@4~JKkn7PX*3OzU{ULE|h+@Ya7Vw=P3s+U;~RtJ-t
zPORIpPGs_`hTogD@-XN2hBrq}4V%~<&HnbI`smi54pZ_H*n+7#so_JrnnLe)jD-9E
z$G44XE^`A9R<3Jt;FfQB_xAYWow#?+jfS=xPc%2}UR>B((cj#n=necy!To!Gqm%!I
zZxdlU!rSlOpRu^b>r(Vk%mAXp7u@vO!I(-AV;{072*pa~2c5m`ee%OaLgM?)7oX)j
zoo2&e|MMmw0VTiz81uiX>V$EGLVe@^v8r4Q%}-c0LZi*%mQ}a@qpBPtE5!e2Rh{6x
zjylTzIbQaLi+$iqRn^M#
z8(-c&vT)bNXIG%Ak~n6Zd3i$+`+^YFt{9W%k(?4%vD_#6WbfyeLRIzn*1zaY1tKGdnJUuawKRR^yJ662$(Kz+*Zp1Yj#^c5%uhH{Or7xozTGNjAS#TkqGV^6#v<@p^xHW*z~U;>txw96XtaV;{r
zCjc)!ZYirHIM3&1ZD=lBr~SeR;nwVrM4Ypl>_Wc-g6--Su|6-MxA;a#bGk;{oiU8
z25Rd4H-s_uD-dHb+L3wB21-Y_qC|CNBtQl1vY2HuI5i#19tqh;@p%z|ZsM}6%^bQH*@3Y7TjNh{kI27D
zFHSt)U~!+(h5kDXq>Khz-pdrbYw8FdG=MmvLI?DBqDTH%$0jgQzLSZBfS%^9BKdP)
zZGO7g>i`=}OU?h-RgQ%SO%C*__|ofN59hW*h6O3wwNz1`7eUhv{vx9FcM^O;c$)_>
z0wbj^k*YX}ZjgoBWhe`_tLsB`mHG1oW5@tOu-OXk;^4V71WyS!R7zLbsp(=ZggnR|
z6*^Ap7MZG?aM$WPEk^0ZrX`vkF)9?RRTe0{`Y+$0s%P$Adhd3gwa%&iY-(y9(k-If
zJgr%PuHf(6iK*A=m7+gKaNuq|(DoZwbbzSA?~bx2rlg{L0hkci2s3*o2&O6mNuTRT
z;C&U88L33^`Q3Q)RKUF}@difKY1Wa-MvT))I|t(orj?-%Z!)yXyd|l)6k6Xq4-81|
z526nCi-V3RV!6TzY6m&U;`kVXq}q&Zp%u*K!;q%NX~LQtOI5
z+TXx)1bwa*Y6<!c7v=wrgLTTGtGyjc>O3h{6R-6kAHd{Cwap1&EFQS1SDH`PgqTps0v8UZfS
z93zFOfsM0f2$mYTTahGQNJD>qZFf3BYmivoBj~eo=D-34-7UEO5I;(ykc3!bx^0Aq
zI3HN*%2l7!kj0>mX1*6!c|=pW-dvrogWn;?4siN@*gd>kQ%*L!rw&3{EA!KJx@^xXinbmJLV>~c_0~8EcrwDgrE|FuFh#H&Bgk=AIk!k<;@V)R}qCOu8P)rbY
z+XcGH;S!4w+I1a|1=z;a`xd&6)K@qaCY4v3@P+NeQTMv{AW<85-Nx_wgWq4Di_B43
zQ`xU;^zZfUBiF?1eUJvPBJqndMy;tIR0xQNh~T9L1N(Vs!r$BZNxq3Dq5)IyY>lNp
zmuECXLz@Lu%XTb(s7=KF00<9-cxtA1tCH1;Yx2I?LJv8Sv#5=HKM`+QtAvLu
zdp)>hq3LlGb0wKw5+FzBOpoi4wPO1Og$ToMFtI98T94`$;2H(+)r=l1L!p$xRl=8Q
zoALqIVWJ`bG3dwX^$n=YhmL3vMzNUJFZ3?0^+5jlAs{c>oTtmBVP62fc&e;`9|b2b
zW+S{a+^52Ox~~*^_ok8VrShI#+?};F7-_e916;u~n&jvy_rW%Pak+Fq?L&|suM>>l
z-f$!f=yTQKoaAJ7plape`tl_S;xA8dMl31l>RuY_FR=Qy8kHX)&(HK9N7^%l@R&p*
zTWm@vD=*S0*fcflpJM7)KKj-cQ|J+9?>ab{1J2P6cL~$RsXT9b_>e)80HxM~s}kI`MBE
ztB(~1R<2wQ97z2!AOXtac~gLo{pa~KGI}TLn^6!g!V3-~lapM}IxNz_YU7=?p+sXJ
z7+Qnc{v9C5acUV_yA-9S;c;qM8kq`{DRK!=*}Z7onJi$yaxEKUNbDtM}${p5vBxb)1zg53f(@eB=^
z9fzGhyT^tXW1$Bt=A}}PXEiXR?GBLZ4sF&l!B8e)j02o%kf0M`6X)RONmh=%FwHB(
zpptOM%blVIR!YfllVR77Zzj=T{&|$o`PR1D{3Y8{iE2Iqt$@3415VLM01G*)AW8rj
z1wcJf>wpS)hn)CL51h{n5^(^u5>>5{JX7nuGJ;jflBvt)yj%yvXQVj4*`@)|S23h+
zK7f{CX9R>mwp);Tz8@d`j)^O)VBKKq=)9U+%qQJgf=ZDiT|NQlXrwC&xYq?@HUJ~2
zl0^V8U!k+`fsn%^zF;DU!$@j%Y%d0Qr@%YYfK%IntP3b74Dli#(39DYsO{TT(Z`G&
z->oIRut(hI6EA-Ojsv82Hata1_(H?~)%dK0mR%{63YNL0-y+A!E7ZVhtHy_
zkt0l;A5&`SoLV?5HE2p91pM6a3v%CX%FaLGhms5G+{`3DcH0l5C2X{>^hHeCWGc$IdX_cdIu2u1O%l9
zU?{;fHnA&&&@CrYlVHi`hz}nVXULdamvknWXk890TTZ-Z1llSv$1g+O`pXH0R#{&*
zykLiAD}6f}@$Y2oaxgmCedzZ*M-4!l;~V6~)EVOu1Hy=7d=y;{^eBk$<)okLBlUb*
z&JVm=9dKWUI=;+d_3Va{JVKhvouJlX(&%r)>-HhKqi{GJ?Vh6=t`hBUkhE>)nCx6wiY)K)%vD;eo8
z0@tdL?jmrZ0@d@D%+P>bJ__<+ZBT##3e>DJMFOP4$49G1ZU@jSRESMv
z&rLG)E&%=HK6;^sMCBmY%D4@GDrj+YQ7K#3Qi3cr#q
z;KXjN00r6b;@PJmzjqcC=*>pwaFG6Uz%V8{3pl$=hTf>Lj@pJwG-;gFnjxm1HUeRUu%0smsYyc*-Z)iy&Z<)L-)T9qJxnYvbSX=
zE$sM}+W0xt)qMothkDt`|F|w4Mx*kq_eMY#Zx8&y#$VyZi%ps0G>kwf&}D)Y(^4J*oB|fR(?G2`I{u4zfz6G!QJ6g>q=wut1+#?PN=*@v3
z$g_A{ZZSXu*ae-;VY{s|%po#1je`KE$W$46CwtDaVe}^Q)Cp~*
z3{tGgFfx0Y-)-{RzRS^ZFuy3%Br~|40N_mnyfhG_UK*R-yGmyV`w=y*wKw&5Z^KwG
zNtgV;@4lA+8qL1~djHpbFLVn{I{vZ?x>yTp8Je*`17jD{`>6{0%`&pD$UJgUQX$1A
zqSj_igxY6dv%s-uwhU2x$b@j6k6qk!x7riGZt!w>YUkm3xr;no&c6*WaaqyI8wOM0
z>z~}=%Ajbjg1B!7OL*_uyk}z5
zougMa@Z!x=!-k4q)Mp#DIFET%UbSAkm~vhnQhoM3BF{9c`crbUxW