chore: add local batch transcription and staging test scripts

Author: hammertoe

runme.sh

@@ -0,0 +1,51 @@
+#!/usr/bin/env bash
+set -euo pipefail
+IDS=(
+  LPtdCCZcOOM
+  df7X19vi3oo
+  v4eK88964Q0
+  J5SQ7bIirHo
+  8CheVcGJgIo
+  INF0vvE7MgE
+  fAzT173IRRg
+  uzKDi93xaa0
+  kzdvMghUNgc
+  ONsxyiKjRHA
+  uxh5Oyq1SAI
+  VKrnJ8ulrgI
+  KTu-MFFamGQ
+  8JahrKBqMtw
+  iyeANTWj-w8
+  aSjd9O94MY0
+  AVT35BrPZIA
+  xKChhsGpcf4
+  WcZ4ZlS5YgI
+  cTn1apqwjFk
+  8ZyEWl2BpDU
+  yYnhpwqZheY
+  PU4vmxf9GwM
+  3mWD_cAFUf4
+  YoIx5icYLR0
+  twui16wvTck
+  Fr_GaEW-ORE
+  z-Fx-ujvLYw
+  _MndwltYWJg
+  UBQGqpv_TN4
+  BWGtHgE2ZOY
+  jWIVADUCoYE
+  t_AdkT9q6z0
+  iQZXLJ5T5yM
+  LTWQaAfcATE
+  p5Kvc2vfmrg
+)
+for vid in "${IDS[@]}"; do
+  echo "=== Transcribing ${vid} ==="
+  python transcribe.py \
+    --video="${vid}" \
+    --segment-minutes 30 \
+    --overlap-minutes 1 \
+    --caption-context \
+    --caption-context-buffer-seconds 30 \
+    --caption-context-max-chars 1200 \
+    --output-file "transcription_output_${vid}.json"
+done

test_staging.py

@@ -0,0 +1,37 @@
+#!/usr/bin/env python3
+import json
+import urllib.request
+
+
+def req(method, url, data=None):
+    b = json.dumps(data).encode() if data else None
+    r = urllib.request.Request(
+        url, data=b, headers={"Content-Type": "application/json"}, method=method
+    )
+    return json.loads(urllib.request.urlopen(r, timeout=90).read().decode())
+
+
+q = "What did ministers say about water management recently?"
+for port, name in [(8013, "green"), (8003, "blue")]:
+    try:
+        t = req("POST", f"http://127.0.0.1:{port}/chat/threads")
+        res = req(
+            "POST",
+            f"http://127.0.0.1:{port}/chat/threads/{t['thread_id']}/messages",
+            {"content": q},
+        )
+        dbg = (res.get("debug") or {}).get("retrieval") or {}
+        print(f"[{name}]")
+        print(f"  edge_count={dbg.get('edge_count')}")
+        print(f"  node_count={dbg.get('node_count')}")
+        print(f"  seed_count={dbg.get('seed_count')}")
+        print(f"  threshold={dbg.get('edge_rank_threshold')}")
+        print(f"  filtered={dbg.get('edges_filtered_by_threshold')}")
+        print(f"  skipped={dbg.get('edge_rank_filter_skipped_no_scores')}")
+        print(f"  sources={len(res.get('sources') or [])}")
+        print(
+            f"  snippet={(res.get('assistant_message', {}).get('content', '')[:150]).replace(chr(10), ' ')}"
+        )
+    except Exception as e:
+        print(f"[{name}] ERROR: {e}")
+    print()