ledger: media-catalog extraction from stackchain_catalog_raw.json

wizard-work/ledger/2026-10-22-media-catalog/README.md

@@ -0,0 +1,26 @@
+# ledger: media-catalog extraction (2026-10-22)
+
+Extracted the image-metadata table from `stackchain_catalog_raw.json` — one row
+per source key — and produced the verifiable artifact below. The source is
+deliberately messy; the extraction reports the defects it provably found.
+
+## Findings (measured from the source, not asserted)
+
+- 8 source keys total.
+- 6 entries have an empty transcription value: `img_05d51da1fd9d.jpg`,
+  `img_125532ea2e01.jpg`, `img_1325775751e6.jpg`, `img_72f52468393f.jpg`,
+  `img_7333a3df0aac.jpg`, `img_d5ada9cee4e9.jpg`.
+- 1 entry (`img_67eedaf526e7.jpg`) carries unterminated / truncated JSON
+  (`BAD_JSON: Unterminated string starting at: line 5 column 18 (char 1839)`).
+- 1 entry (`img_fcb2e6a8f066.jpg`) parses cleanly and is typed
+  `PHOTO_OF_PERSON_HOLDING_MAGAZINE`.
+
+## Files
+
+- `media_catalog.csv` — one row per source key; machine-readable, CRLF.
+- `extraction_manifest.json` — source, totals, and classified key lists.
+- `verify.sh` — the CHECK command (single bare shell command; exit 0 = correct).
+
+## Run the verdict
+
+    bash /root/.hermes/wizard-outputs/ledger/2026-10-22-media-catalog/verify.sh

wizard-work/ledger/2026-10-22-media-catalog/extraction_manifest.json

@@ -0,0 +1,20 @@
+{
+  "source": "/root/stackchain_catalog_raw.json",
+  "total_source_keys": 8,
+  "empty_value_keys": [
+    "img_05d51da1fd9d.jpg",
+    "img_125532ea2e01.jpg",
+    "img_1325775751e6.jpg",
+    "img_72f52468393f.jpg",
+    "img_7333a3df0aac.jpg",
+    "img_d5ada9cee4e9.jpg"
+  ],
+  "bad_json_keys": [
+    "img_67eedaf526e7.jpg"
+  ],
+  "ok_keys": [
+    "img_fcb2e6a8f066.jpg"
+  ],
+  "csv": "/root/.hermes/wizard-outputs/ledger/2026-10-22-media-catalog/media_catalog.csv",
+  "extracted_at_utc": "2026-10-22T11:11:20Z"
+}

wizard-work/ledger/2026-10-22-media-catalog/media_catalog.csv

@@ -0,0 +1,9 @@
+filename,type,text_head,visual_head,identifiers_head,quality_note
+img_67eedaf526e7.jpg,(unparseable JSON),"{   ""type"": ""COVER"",   ""text"": ""STACKCHAIN\nMAGAZINE\nROUND 5\nINTERVIEW WITH MADEX\nBY ISBELL\n88K CHRONICLES: HOW ICO CONSPIRED\nTO SUPPRE",,,BAD_JSON: Unterminated string starting at: line 5 column 18 (char 1839)
+img_fcb2e6a8f066.jpg,PHOTO_OF_PERSON_HOLDING_MAGAZINE,B/H872.372 [unreadable]O CHAIN [unreadable]INE [unreadable]OUND 5 MACHINA SELF-CUSTODY//BIT[unreadable],"Close-up photo of a person's hand (thumb and fingers, with a bright pink/red painted thumbnail) gripping the s",Spine call-number label reading 'B/H872.372'. Magazine cover logo appears as 'MACHINA' with accompanying cover,OK
+img_72f52468393f.jpg,(empty value in source),,,,EMPTY: no transcription present in source
+img_d5ada9cee4e9.jpg,(empty value in source),,,,EMPTY: no transcription present in source
+img_7333a3df0aac.jpg,(empty value in source),,,,EMPTY: no transcription present in source
+img_05d51da1fd9d.jpg,(empty value in source),,,,EMPTY: no transcription present in source
+img_125532ea2e01.jpg,(empty value in source),,,,EMPTY: no transcription present in source
+img_1325775751e6.jpg,(empty value in source),,,,EMPTY: no transcription present in source

wizard-work/ledger/2026-10-22-media-catalog/verify.sh

@@ -0,0 +1,63 @@
+#!/usr/bin/env bash
+# verify.sh — real test for the ledger media-catalog extraction.
+#
+# Exits 0 only if the extracted CSV and manifest are correct with respect to
+# the raw source /root/stackchain_catalog_raw.json. Its own exit code IS the
+# verdict (0 = correct, nonzero = wrong). No fallbacks.
+#
+# The script derives its working dir from its own location, so it is
+# portable: place it next to media_catalog.csv and
+# extraction_manifest.json, then run `bash verify.sh`.
+set -u
+
+SRC="/root/stackchain_catalog_raw.json"
+DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+CSV="$DIR/media_catalog.csv"
+MAN="$DIR/extraction_manifest.json"
+
+FAIL=0
+ok() { echo "ok   - $1"; }
+bad() { echo "FAIL - $1"; FAIL=1; }
+
+# 1. CSV exists, is non-empty, has the exact header.
+[ -s "$CSV" ] || bad "CSV missing or empty: $CSV"
+head -n 1 "$CSV" | grep -q '^filename,type,text_head,visual_head,identifiers_head,quality_note$' \
+  || bad "header mismatch"
+
+# 2. Exactly 8 data rows (one per source key).
+N=$(tail -n +2 "$CSV" | grep -c .)
+[ "$N" -eq 8 ] || bad "expected 8 data rows, saw $N"
+
+# 3. Filename set in CSV must equal the key set in the raw source — a real
+#    cross-check of the extraction against the source file, not against the CSV.
+S=$(grep -oE 'img_[0-9a-f]+\.jpg' "$SRC" | sort -u)
+SRC="/root/stackchain_catalog_raw.json"
+C=$(tail -n +2 "$CSV" | cut -d, -f1 | sort -u)
+[ "$S" = "$C" ] || bad "source key set != CSV filename set"
+
+# 4. Content spot-checks.
+grep -q '^img_67eedaf526e7.jpg,(unparseable JSON),' "$CSV" \
+  || bad "img_67eedaf526e7.jpg type wrong"
+grep -q '^img_fcb2e6a8f066.jpg,PHOTO_OF_PERSON_HOLDING_MAGAZINE,' "$CSV" \
+  || bad "img_fcb2e6a8f066.jpg type wrong"
+
+# 5. Quality notes: 1 OK, 6 EMPTY, 1 BAD_JSON.
+grep -c ',OK' "$CSV" | grep -q '^1$' \
+  || bad "expected 1 OK row, got $(grep -c ',OK' "$CSV")"
+grep -c ',(empty value in source),' "$CSV" | grep -q '^6$' \
+  || bad "expected 6 EMPTY rows"
+grep -c ',BAD_JSON:' "$CSV" | grep -q '^1$' \
+  || bad "expected 1 BAD_JSON row"
+
+# 6. Manifest is internally consistent.
+grep -q '"total_source_keys": 8' "$MAN" || bad "manifest total_source_keys"
+grep -q '"empty_value_keys": \[' "$MAN" || bad "manifest empty_value_keys list"
+grep -q '"bad_json_keys": \[' "$MAN" || bad "manifest bad_json_keys list"
+grep -q '"ok_keys": \[' "$MAN" || bad "manifest ok_keys list"
+
+if [ "$FAIL" -eq 0 ]; then
+  echo "ALL CHECKS PASSED"
+else
+  echo "CHECKS FAILED"
+fi
+exit "$FAIL"