cellarbrain 0.2.2__tar.gz → 0.2.4__tar.gz

{cellarbrain-0.2.2/src/cellarbrain.egg-info → cellarbrain-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cellarbrain
-Version: 0.2.2
+Version: 0.2.4
 Summary: AI sommelier for your wine cellar — ETL pipeline, DuckDB query layer, Markdown dossiers, and MCP server for wine cellar CSV exports
 Author-email: Urban Busslinger <urbanb@me.com>
 License-Expression: MIT

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "cellarbrain"
-version = "0.2.2"
+version = "0.2.4"
 description = "AI sommelier for your wine cellar — ETL pipeline, DuckDB query layer, Markdown dossiers, and MCP server for wine cellar CSV exports"
 requires-python = ">=3.11"
 license = "MIT"

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/cli.py

@@ -1349,6 +1349,9 @@ def _cmd_ingest(args: argparse.Namespace, settings: Settings) -> None:
 
     if args.once:
         count = poll_once(config, settings, dry_run=args.dry_run)
+        if count < 0:
+            print(f"Failed {-count} batch(es) (ETL error — messages left unprocessed).")
+            sys.exit(1)
         print(f"Processed {count} batch(es).")
         sys.exit(0)
 

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/email_poll/__init__.py

@@ -58,7 +58,8 @@ def poll_once(
 ) -> int:
     """Execute a single poll cycle.
 
-    Returns the number of batches successfully processed (0 or more).
+    Returns the number of batches successfully processed (0 or more),
+    or a negative number indicating how many batches failed ETL.
     """
     from .credentials import resolve_credentials
     from .etl_runner import run_etl
@@ -95,11 +96,26 @@ def poll_once(
         logger.info("Found %d new messages", len(uids))
 
         # Fetch and parse
-        fetched = client.fetch_messages(uids, config.expected_files)
+        fetched = client.fetch_messages(uids, config.expected_files, max_attachment_bytes=config.max_attachment_bytes)
         if not fetched:
             logger.info("No messages with valid attachments")
             return 0
 
+        # Application-level sender whitelist (defence-in-depth)
+        if config.sender_whitelist:
+            whitelist = {s.lower() for s in config.sender_whitelist}
+            original_count = len(fetched)
+            fetched = [(em, data) for em, data in fetched if em.sender in whitelist]
+            rejected = original_count - len(fetched)
+            if rejected:
+                logger.warning(
+                    "Rejected %d message(s) from non-whitelisted senders",
+                    rejected,
+                )
+            if not fetched:
+                logger.info("No messages from whitelisted senders")
+                return 0
+
         # Build EmailMessage list and attachment map
         messages = [em for em, _ in fetched]
         attachment_map: dict[int, tuple[str, bytes]] = {em.uid: (em.filename, data) for em, data in fetched}
@@ -111,6 +127,7 @@ def poll_once(
             return 0
 
         processed = 0
+        failed = 0
         for batch in batches:
             logger.info(
                 "Batch detected — %s",
@@ -141,11 +158,18 @@ def poll_once(
                 output_dir,
                 config_path,
                 expected_files=config.expected_files,
+                timeout=config.etl_timeout,
             )
             if exit_code != 0:
-                logger.error("ETL failed (exit %d)", exit_code)
+                logger.error(
+                    "ETL failed (exit %d) — leaving messages unprocessed (UIDs: %s)",
+                    exit_code,
+                    list(batch.uids),
+                )
+                failed += 1
+                continue
 
-            # Mark as processed (regardless of ETL outcome)
+            # Mark as processed only on successful ETL
             batch_uids = list(batch.uids)
             if config.processed_action == "move":
                 client.move_messages(batch_uids, config.processed_folder)
@@ -161,6 +185,8 @@ def poll_once(
 
             processed += 1
 
+    if failed:
+        return -failed
     return processed
 
 
@@ -176,7 +202,7 @@ class IngestDaemon:
         self.config = config
         self.settings = settings
         self._base_interval = config.poll_interval
-        self._max_interval = 600  # 10 minutes
+        self._max_interval = config.max_backoff_interval
         self._current_interval = config.poll_interval
 
     def run(self, *, dry_run: bool = False) -> None:
@@ -195,9 +221,17 @@ class IngestDaemon:
         while True:
             try:
                 count = poll_once(self.config, self.settings, dry_run=dry_run)
-                if count > 0:
+                if count < 0:
+                    logger.error("ETL failed for %d batch(es) — will retry next cycle", -count)
+                    self._current_interval = min(
+                        self._current_interval * 2,
+                        self._max_interval,
+                    )
+                elif count > 0:
                     logger.info("Processed %d batch(es)", count)
-                self._current_interval = self._base_interval
+                    self._current_interval = self._base_interval
+                else:
+                    self._current_interval = self._base_interval
             except ValueError:
                 # Credential / config errors — fatal, stop daemon
                 raise

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/email_poll/etl_runner.py

@@ -3,14 +3,13 @@
 from __future__ import annotations
 
 import logging
+import os
 import subprocess
 import sys
 from pathlib import Path
 
 logger = logging.getLogger(__name__)
 
-_ETL_TIMEOUT = 300  # seconds
-
 
 def run_etl(
     raw_dir: Path,
@@ -22,6 +21,7 @@ def run_etl(
         "export-bottles-stored.csv",
         "export-bottles-gone.csv",
     ),
+    timeout: int = 300,
 ) -> tuple[int, str]:
     """Run ``cellarbrain etl`` as a subprocess.
 
@@ -36,6 +36,8 @@ def run_etl(
     expected_files:
         Filenames to pass to the ETL command (in positional order:
         wines, bottles-stored, bottles-gone).
+    timeout:
+        Seconds before the ETL subprocess is killed.
 
     Returns
     -------
@@ -58,11 +60,13 @@ def run_etl(
 
     logger.info("Running ETL: %s", " ".join(cmd))
     try:
+        env = {**os.environ, "PYTHONUTF8": "1", "PYTHONIOENCODING": "utf-8"}
         result = subprocess.run(
             cmd,
             capture_output=True,
             text=True,
-            timeout=_ETL_TIMEOUT,
+            timeout=timeout,
+            env=env,
         )
         output = result.stdout + result.stderr
         if result.returncode == 0:
@@ -71,5 +75,5 @@ def run_etl(
             logger.error("ETL failed (exit %d): %s", result.returncode, output)
         return result.returncode, output
     except subprocess.TimeoutExpired:
-        logger.error("ETL timed out after %d seconds", _ETL_TIMEOUT)
-        return -1, f"ETL timed out after {_ETL_TIMEOUT} seconds"
+        logger.error("ETL timed out after %d seconds", timeout)
+        return -1, f"ETL timed out after {timeout} seconds"

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/email_poll/grouping.py

@@ -21,6 +21,7 @@ class EmailMessage:
     date: datetime
     filename: str
     size: int
+    sender: str = ""
 
 
 @dataclass(frozen=True)

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/email_poll/imap.py

@@ -9,6 +9,7 @@ from __future__ import annotations
 
 import email
 import email.policy
+import email.utils
 import logging
 from datetime import UTC, datetime
 from types import TracebackType
@@ -91,26 +92,35 @@ class ImapClient:
         self,
         uids: list[int],
         expected_files: tuple[str, ...] | list[str],
+        *,
+        max_attachment_bytes: int = 0,
     ) -> list[tuple[EmailMessage, bytes]]:
         """Fetch messages and extract single-attachment metadata + data.
 
         Only messages with exactly one attachment whose filename is in
         *expected_files* are returned.  Others are silently skipped.
 
+        Parameters
+        ----------
+        max_attachment_bytes:
+            If > 0, skip attachments exceeding this size (bytes).
+
         Returns list of ``(EmailMessage, attachment_bytes)`` tuples.
         """
         if not uids:
             return []
 
         results: list[tuple[EmailMessage, bytes]] = []
-        raw_responses = self._client.fetch(uids, ["RFC822", "INTERNALDATE"])
+        raw_responses = self._client.fetch(uids, ["BODY.PEEK[]", "INTERNALDATE"])
 
         for uid, data in raw_responses.items():
             internal_date = data.get(b"INTERNALDATE")
             if internal_date is None:
                 internal_date = datetime.now(UTC)
 
-            rfc822 = data.get(b"RFC822", b"")
+            rfc822 = data.get(b"BODY[]") or data.get(b"RFC822", b"")
+            if not rfc822:
+                continue
             msg = email.message_from_bytes(rfc822, policy=email.policy.default)
 
             attachments = _extract_attachments(msg)
@@ -121,11 +131,25 @@ class ImapClient:
             if filename not in expected_files:
                 continue
 
+            if max_attachment_bytes and len(payload) > max_attachment_bytes:
+                logger.warning(
+                    "Attachment %s (%d bytes) exceeds limit — skipping UID %d",
+                    filename,
+                    len(payload),
+                    uid,
+                )
+                continue
+
+            # Extract sender from From: header
+            from_header = msg.get("From", "")
+            _, sender_addr = email.utils.parseaddr(from_header)
+
             em = EmailMessage(
                 uid=int(uid),
                 date=internal_date,
                 filename=filename,
                 size=len(payload),
+                sender=sender_addr.lower(),
             )
             results.append((em, payload))
 

{cellarbrain-0.2.2 → cellarbrain-0.2.4}/src/cellarbrain/settings.py

@@ -242,6 +242,7 @@ class IngestConfig:
     mailbox: str = "INBOX"
     subject_filter: str = "[VinoCell] CSV file"
     sender_filter: str = ""
+    sender_whitelist: tuple[str, ...] = ()
     poll_interval: int = 60
     batch_window: int = 300
     expected_files: tuple[str, ...] = (
@@ -251,6 +252,9 @@ class IngestConfig:
     )
     processed_action: str = "flag"
     processed_folder: str = "VinoCell/Processed"
+    etl_timeout: int = 300
+    max_backoff_interval: int = 600
+    max_attachment_bytes: int = 10_485_760
 
 
 # ---------------------------------------------------------------------------
@@ -919,12 +923,14 @@ def load_settings(
         _validate_keys("dashboard", dashboard_raw, DashboardConfig)
     dashboard = DashboardConfig(**dashboard_raw) if dashboard_raw else DashboardConfig()
 
-    # Ingest — scalar config with tuple conversion for expected_files
+    # Ingest — scalar config with tuple conversion for expected_files/sender_whitelist
     ingest_raw = raw.get("ingest", {})
     if ingest_raw:
         ingest_kw: dict = dict(ingest_raw)
         if "expected_files" in ingest_kw:
             ingest_kw["expected_files"] = tuple(ingest_kw["expected_files"])
+        if "sender_whitelist" in ingest_kw:
+            ingest_kw["sender_whitelist"] = tuple(ingest_kw["sender_whitelist"])
         _validate_keys("ingest", ingest_kw, IngestConfig)
         ingest = IngestConfig(**ingest_kw)
     else:

{cellarbrain-0.2.2 → cellarbrain-0.2.4/src/cellarbrain.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cellarbrain
-Version: 0.2.2
+Version: 0.2.4
 Summary: AI sommelier for your wine cellar — ETL pipeline, DuckDB query layer, Markdown dossiers, and MCP server for wine cellar CSV exports
 Author-email: Urban Busslinger <urbanb@me.com>
 License-Expression: MIT