stock-weekly-report 0.2.1 → 0.2.2

package/fetch_episodes.py

@@ -6,7 +6,10 @@ Fetches RSS feeds from Soundon, finds episodes published within the
 configured lookback window, and downloads the audio files.
 
 Output structure:
-  {parent_folder}/audio/{YYYYMMDD}-{YYYYMMDD}/{program_name}_{YYYYMMDD}.ext
+  {parent_folder}/audio/{program_name}/{program_name}_{YYYYMMDD}.ext
+
+Each speaker has a persistent folder. Downloads are skipped if the file
+already exists (checked by filename / date), so re-runs are safe.
 """
 
 import os
@@ -144,11 +147,10 @@ def fetch_and_download(config: dict, folder_name: str | None = None) -> None:
         end_date   = date(int(parts[1][:4]), int(parts[1][4:6]), int(parts[1][6:8]))
 
     parent_folder = Path(config["parent_folder"])
-    audio_dir = parent_folder / "audio" / run_folder
-    audio_dir.mkdir(parents=True, exist_ok=True)
+    audio_root = parent_folder / "audio"
 
     print(f"Date range  : {start_date} → {end_date}")
-    print(f"Audio folder: {audio_dir}")
+    print(f"Audio root  : {audio_root}")
     print()
 
     total_downloaded = 0
@@ -157,6 +159,9 @@ def fetch_and_download(config: dict, folder_name: str | None = None) -> None:
         program_name = feed_cfg["name"]
         feed_url = feed_cfg["url"]
 
+        speaker_dir = audio_root / program_name
+        speaker_dir.mkdir(parents=True, exist_ok=True)
+
         print(f"[{program_name}] Fetching feed …")
         parsed = feedparser.parse(feed_url)
 
@@ -180,7 +185,7 @@ def fetch_and_download(config: dict, folder_name: str | None = None) -> None:
             ext = url_extension(audio_url)
             date_str = pub_date.strftime("%Y%m%d")
             filename = f"{program_name}_{date_str}{ext}"
-            dest = audio_dir / filename
+            dest = speaker_dir / filename
 
             if dest.exists():
                 print(f"  SKIP (already downloaded): {filename}")
@@ -203,7 +208,7 @@ def fetch_and_download(config: dict, folder_name: str | None = None) -> None:
             print(f"  No new episodes in the past {lookback_days} days.")
         print()
 
-    print(f"Done. {total_downloaded} file(s) newly downloaded → {audio_dir}")
+    print(f"Done. {total_downloaded} file(s) newly downloaded → {audio_root}")
 
 
 def run(config_path: str = "config.yaml") -> None:

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "stock-weekly-report",
-  "version": "0.2.1",
+  "version": "0.2.2",
   "description": "Stock weekly podcast report pipeline — CLI and MCP server",
   "bin": {
     "swr": "bin/swr.js",

package/pipeline.py

@@ -79,15 +79,24 @@ def validate_audio_files(config: dict, folder_name: str) -> bool:
       - Returns False (abort) only when zero usable audio files remain after cleanup.
     """
     banner("GUARD — Audio File Integrity")
-    audio_dir = Path(config["parent_folder"]) / "audio" / folder_name
+    audio_root = Path(config["parent_folder"]) / "audio"
+    parts = folder_name.split("-")
+    start_str, end_str = parts[0], parts[1]
 
     audio_files = []
-    for ext in SUPPORTED_AUDIO_EXTS:
-        audio_files.extend(audio_dir.glob(f"*{ext}"))
+    if audio_root.exists():
+        for speaker_dir in sorted(audio_root.iterdir()):
+            if not speaker_dir.is_dir():
+                continue
+            for ext in SUPPORTED_AUDIO_EXTS:
+                for f in speaker_dir.glob(f"*{ext}"):
+                    date_str = f.stem.split("_")[-1]
+                    if len(date_str) == 8 and start_str <= date_str <= end_str:
+                        audio_files.append(f)
     audio_files = sorted(audio_files)
 
     if not audio_files:
-        print(f"  ERROR: No audio files found in {audio_dir}")
+        print(f"  ERROR: No audio files found in {audio_root}")
         return False
 
     usable = 0
@@ -244,15 +253,51 @@ def _cleanup_data_dir(data_root: Path, label: str,
         print(f"\n  Cleaned up {removed_folders} folder(s).")
 
 
+def _cleanup_audio_by_speaker(audio_root: Path, months: int) -> None:
+    """Delete audio files older than `months` from per-speaker subdirectories."""
+    if months <= 0:
+        print(f"  Audio: retention = 0 (keep forever), skipping.")
+        return
+
+    banner(f"CLEANUP — Audio (keep {months} month{'s' if months != 1 else ''})")
+
+    if not audio_root.exists():
+        print(f"  Directory not found: {audio_root}")
+        return
+
+    cutoff = _cutoff_date(months)
+    print(f"  Cutoff date : {cutoff}  (deleting files published before this date)")
+
+    removed = 0
+    for speaker_dir in sorted(audio_root.iterdir()):
+        if not speaker_dir.is_dir():
+            continue
+        for ext in SUPPORTED_AUDIO_EXTS:
+            for f in speaker_dir.glob(f"*{ext}"):
+                date_str = f.stem.split("_")[-1]
+                try:
+                    file_date = datetime.strptime(date_str, "%Y%m%d").date()
+                except ValueError:
+                    continue
+                if file_date < cutoff:
+                    size_mb = f.stat().st_size / (1024 * 1024)
+                    print(f"  Deleting ({size_mb:.1f} MB): {speaker_dir.name}/{f.name}")
+                    f.unlink()
+                    removed += 1
+
+    if removed == 0:
+        print("  No old audio files to remove.")
+    else:
+        print(f"  Removed {removed} file(s).")
+
+
 def cleanup_old_data(config: dict) -> bool:
     """Clean up old audio, transcript, and report files per retention config."""
     retention = config.get("retention", {})
     parent    = Path(config["parent_folder"])
 
-    _cleanup_data_dir(
+    _cleanup_audio_by_speaker(
         parent / "audio",
-        "Audio",
-        SUPPORTED_AUDIO_EXTS,
         int(retention.get("audio_months", 3)),
     )
     _cleanup_data_dir(
@@ -340,7 +385,7 @@ def main() -> None:
         results["fetch"] = "skipped"
 
     # ── Guard: Audio integrity ───────────────────────────────────────
-    if not args.skip_fetch and not args.skip_transcribe:
+    if not args.skip_transcribe:
         ok = validate_audio_files(config, folder_name)
         results["audio_check"] = ok
         if not ok:

package/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "stock-weekly-report"
-version = "0.2.1"
+version = "0.2.2"
 requires-python = ">=3.10"
 authors = [
     { name = "Chang Yu Chuan", email = "changyuchuanmicron@gmail.com" },

package/send_report.py

@@ -439,7 +439,7 @@ def save_report(config: dict, folder_name: str, body: str) -> Path:
     """Save the report text to disk and return the file path."""
     report_dir = Path(config["parent_folder"]) / "reports" / folder_name
     report_dir.mkdir(parents=True, exist_ok=True)
-    report_path = report_dir / "weekly_report.txt"
+    report_path = report_dir / f"weekly_report_{folder_name}.txt"
     report_path.write_text(body, encoding="utf-8")
     print(f"  Report saved to: {report_path}")
     return report_path

package/transcribe.py

@@ -15,6 +15,9 @@ Usage:
   python transcribe.py --folder 20260218-20260225
   python transcribe.py --config my_config.yaml --folder 20260218-20260225
 
+Input structure:
+  {parent_folder}/audio/{program_name}/{program_name}_{YYYYMMDD}.ext
+
 Output structure:
   {parent_folder}/transcripts/{YYYYMMDD}-{YYYYMMDD}/{stem}.txt
 """
@@ -51,10 +54,26 @@ def default_folder_name(lookback_days: int) -> str:
     return f"{start.strftime('%Y%m%d')}-{today.strftime('%Y%m%d')}"
 
 
-def find_audio_files(audio_dir: Path) -> list[Path]:
+def find_audio_files_for_run(audio_root: Path, folder_name: str) -> list[Path]:
+    """Collect audio files across per-speaker subdirs whose date falls in the run window.
+
+    Expects filenames of the form {speaker}_{YYYYMMDD}.ext so the date can be
+    extracted from the stem suffix.
+    """
+    parts = folder_name.split("-")
+    start_str, end_str = parts[0], parts[1]
+
     files = []
-    for ext in SUPPORTED_AUDIO_EXTS:
-        files.extend(audio_dir.glob(f"*{ext}"))
+    if not audio_root.exists():
+        return files
+    for speaker_dir in sorted(audio_root.iterdir()):
+        if not speaker_dir.is_dir():
+            continue
+        for ext in SUPPORTED_AUDIO_EXTS:
+            for f in speaker_dir.glob(f"*{ext}"):
+                date_str = f.stem.split("_")[-1]
+                if len(date_str) == 8 and start_str <= date_str <= end_str:
+                    files.append(f)
     return sorted(files)
 
 
@@ -150,18 +169,14 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
     from faster_whisper import WhisperModel
 
     parent_folder = Path(config["parent_folder"])
-    audio_dir = parent_folder / "audio" / folder_name
+    audio_root = parent_folder / "audio"
     transcript_dir = parent_folder / "transcripts" / folder_name
 
-    if not audio_dir.exists():
-        print(f"ERROR: Audio directory not found: {audio_dir}")
-        sys.exit(1)
-
     transcript_dir.mkdir(parents=True, exist_ok=True)
 
-    audio_files = find_audio_files(audio_dir)
+    audio_files = find_audio_files_for_run(audio_root, folder_name)
     if not audio_files:
-        print(f"No audio files found in {audio_dir}")
+        print(f"No audio files found in {audio_root} for run window {folder_name}")
         return
 
     model_name = config.get("whisper_model", "medium")
@@ -171,7 +186,7 @@ def transcribe_folder(config: dict, folder_name: str) -> None:
     print(f"Whisper model   : {model_name}  (faster-whisper / CTranslate2)")
     print(f"Compute type    : {compute_type}")
     print(f"Language hint   : {language}")
-    print(f"Audio folder    : {audio_dir}")
+    print(f"Audio root      : {audio_root}")
     print(f"Transcript dir  : {transcript_dir}")
     print(f"Files to process: {len(audio_files)}")
     print(f"Max retries     : {MAX_RETRIES}")