daily/aashto: track per-year Crash.csv, sync cc2mc2mn.json dep hash

Root-cause the aashto-stage rebuild-loop that broke today's daily.
Three latent brittle spots collided when my earlier pull-list expansion
forced DVX to walk upstream instead of stopping at the target stage:

1. **Per-year `Crash.csv` was on disk only (not DVC-tracked)** for
   2023/2024/2025. `dvx add` all three (~2 GB total) and push. Each
   `Crash.csv.dvc` md5 matches the corresponding `crashes.parquet.dvc`
   dep hash, so downstream stages see the dep as satisfied without
   needing `normalize.py` to re-run.

2. **Daily didn't pull `Crash.csv.dvc`** even though it's an upstream
   dep of the per-year `crashes.parquet` stage — pulls added.

3. **`cc2mc2mn.json` was updated 2026-07-04 (`ca5cef2b69a`) without
   refreshing the aashto-chain .dvc dep hashes**. Bumped in the four
   affected files:
     - `aashto_combined_crashes.parquet.dvc`
     - `aashto_supplemented_occupants.parquet.dvc`
     - `aashto_supplemented_pedestrians.parquet.dvc`
     - `aashto_supplemented_vehicles.parquet.dvc`
   Output md5 is unchanged (verified on disk) — this is metadata sync,
   no recompute.

Also expand daily.yml comment above the per-year pulls to explain the
`crashes.parquet.dvc` + `Crash.csv.dvc` pairing (was #127-adjacent).

Prior daily runs (07-05 → 07-07) succeeded despite this staleness only
because DVX's plan stopped at the direct target (`njsp_njdot_match`,
already up-to-date). Today's plan walks 4 levels due to more per-year
outputs joining the pull list, exposing the stale chain.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

Author: ryan-williams

.github/workflows/daily.yml

@@ -79,16 +79,18 @@ jobs:
           dvx pull njdot/data/2023/vehicles.parquet.dvc
           dvx pull njdot/data/2024/vehicles.parquet.dvc
           dvx pull njdot/data/2025/vehicles.parquet.dvc
-          # Per-year AASHTO `crashes.parquet` — same rationale as persons /
-          # vehicles above. Consumed transitively by
-          # `aashto_combined_crashes.parquet` (which is pulled below). Without
-          # these on disk, DVX sees the intermediate stage's output-fresh
-          # check pass but the *upstream* per-year stage's output missing,
-          # tries to re-run `normalize.py`, and hits the shebang-not-found
-          # path (see #127-adjacent notes).
+          # Per-year AASHTO `crashes.parquet` (produced by `normalize.py` from
+          # the corresponding `Crash.csv`). Same rationale as the persons /
+          # vehicles pulls above — the outputs on disk let DVX report
+          # up-to-date without traversing into the per-year stage. And
+          # `Crash.csv` itself must be pulled too, otherwise the stage's dep
+          # is missing → DVX still tries to re-run.
           dvx pull njdot/data/2023/crashes.parquet.dvc
           dvx pull njdot/data/2024/crashes.parquet.dvc
           dvx pull njdot/data/2025/crashes.parquet.dvc
+          dvx pull njdot/data/2023/Crash.csv.dvc
+          dvx pull njdot/data/2024/Crash.csv.dvc
+          dvx pull njdot/data/2025/Crash.csv.dvc
           dvx pull njsp/data/crash-log.parquet.dvc
           # NJSP outputs of `njsp update_pqts` (consumed by harmonize, csvs,
           # projections, match, three-way-match, d1-import).

njdot/data/2023/Crash.csv.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: df2388790112b398a553c0f67a0cf173
+  size: 650972064
+  hash: md5
+  path: Crash.csv

njdot/data/2024/Crash.csv.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 5dba02a6dcf980f5215499f7a32da89b
+  size: 714975929
+  hash: md5
+  path: Crash.csv

njdot/data/2025/Crash.csv.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: cd50e08bf594eb03a4f2aadeffd2564a
+  size: 713333114
+  hash: md5
+  path: Crash.csv

njdot/data/aashto_combined_crashes.parquet.dvc

@@ -12,4 +12,4 @@ meta:
       /njdot/data/2023/crashes.parquet: 23abc6794e13d6d8b035dfe866de86b9
       /njdot/data/2024/crashes.parquet: fd7cdf13806203af5ddce84ca94d3e23
       /njdot/data/2025/crashes.parquet: 26e6bc7f841bbaa6b5878e57771869dd
-      /www/public/njdot/cc2mc2mn.json: 2c12563c5ea3a3fdd443fc14610822e8
+      /www/public/njdot/cc2mc2mn.json: 21626cd6b5ed02264335837141b86984

njdot/data/aashto_supplemented_occupants.parquet.dvc

@@ -7,7 +7,8 @@ meta:
   computation:
     cmd: njdot aashto persons -y 2023,2024,2025
     git_deps:
-      /njdot/aashto/to_njdot_persons.py: 21b6405b82bcf6c362c6f73052168a11f5896d48
+      /njdot/aashto/to_njdot_persons.py: 
+        21b6405b82bcf6c362c6f73052168a11f5896d48
       /njdot/aashto/to_njdot_schema.py: 358a48b822f7e461f9d2ebef97eac0bbff04958f
     deps:
       /njdot/data/2023/persons.parquet: 72625f5a685ceb7489789e886f4f3171
@@ -16,7 +17,7 @@ meta:
       /njdot/data/2024/crashes.parquet: fd7cdf13806203af5ddce84ca94d3e23
       /njdot/data/2025/persons.parquet: 2c311b054ff1f7fbb27c0c099ae0486b
       /njdot/data/2025/crashes.parquet: 26e6bc7f841bbaa6b5878e57771869dd
-      /www/public/njdot/cc2mc2mn.json: 2c12563c5ea3a3fdd443fc14610822e8
+      /www/public/njdot/cc2mc2mn.json: 21626cd6b5ed02264335837141b86984
     note: |
       AASHTO persons -> DOTr-style occupants supplement for 2023-2025.
       "Ghost-Driver" placeholder rows (Person Type=Driver + blank

njdot/data/aashto_supplemented_pedestrians.parquet.dvc

@@ -7,7 +7,8 @@ meta:
   computation:
     cmd: njdot aashto persons -y 2023,2024,2025
     git_deps:
-      /njdot/aashto/to_njdot_persons.py: 21b6405b82bcf6c362c6f73052168a11f5896d48
+      /njdot/aashto/to_njdot_persons.py: 
+        21b6405b82bcf6c362c6f73052168a11f5896d48
       /njdot/aashto/to_njdot_schema.py: 358a48b822f7e461f9d2ebef97eac0bbff04958f
     deps:
       /njdot/data/2023/persons.parquet: 72625f5a685ceb7489789e886f4f3171
@@ -16,7 +17,7 @@ meta:
       /njdot/data/2024/crashes.parquet: fd7cdf13806203af5ddce84ca94d3e23
       /njdot/data/2025/persons.parquet: 2c311b054ff1f7fbb27c0c099ae0486b
       /njdot/data/2025/crashes.parquet: 26e6bc7f841bbaa6b5878e57771869dd
-      /www/public/njdot/cc2mc2mn.json: 2c12563c5ea3a3fdd443fc14610822e8
+      /www/public/njdot/cc2mc2mn.json: 21626cd6b5ed02264335837141b86984
     note: |
       AASHTO persons → DOTr-style pedestrians supplement for 2023-2025.
       Both pedestrians and cyclists are co-tracked here (`cyclist` bool

njdot/data/aashto_supplemented_vehicles.parquet.dvc

@@ -7,7 +7,8 @@ meta:
   computation:
     cmd: njdot aashto vehicles -y 2023,2024,2025
     git_deps:
-      /njdot/aashto/to_njdot_vehicles.py: 6dbcc982eafe7f2d1bbdbbe2ee8391d96a1bba12
+      /njdot/aashto/to_njdot_vehicles.py: 
+        6dbcc982eafe7f2d1bbdbbe2ee8391d96a1bba12
       /njdot/aashto/to_njdot_schema.py: 358a48b822f7e461f9d2ebef97eac0bbff04958f
     deps:
       /njdot/data/2023/vehicles.parquet: 0203a9b9496e90c31c82fd1940bdfd3e
@@ -16,7 +17,7 @@ meta:
       /njdot/data/2024/crashes.parquet: fd7cdf13806203af5ddce84ca94d3e23
       /njdot/data/2025/vehicles.parquet: b2e6d80ce5275d8165494af61733eeef
       /njdot/data/2025/crashes.parquet: 26e6bc7f841bbaa6b5878e57771869dd
-      /www/public/njdot/cc2mc2mn.json: 2c12563c5ea3a3fdd443fc14610822e8
+      /www/public/njdot/cc2mc2mn.json: 21626cd6b5ed02264335837141b86984
     note: |
       AASHTO vehicles -> DOTr-style vehicles supplement for 2023-2025.
       Provides per-vehicle Damage (`vdn/vdm/vdo/vdx/vdu`) for AASHTO