Address PR #112 review: CollectedExportEntry, shared read_last_export_ms, engine tests

Author: clean6378-max-it

api/export_api.py

@@ -18,9 +18,8 @@
 from flask import Blueprint, Response, request
 
 from api.flask_config import exclusion_rules, json_response
-from services.export_engine import collect_export_entries
+from services.export_engine import collect_export_entries, read_last_export_ms
 from services.workspace_db import global_storage_db_path
-from utils.path_helpers import to_epoch_ms
 from utils.workspace_path import resolve_workspace_path
 
 bp = Blueprint("export_api", __name__)
@@ -67,15 +66,6 @@ def _save_export_state(count: int) -> None:
         json.dump(state, f, indent=2)
 
 
-def _read_last_export_ms(since: Literal["all", "last"]) -> int:
-    if since != "last":
-        return 0
-    ts = _get_export_state().get("lastExportTime")
-    if ts:
-        return to_epoch_ms(ts)
-    return 0
-
-
 @bp.route("/api/export/state")
 def get_export_state() -> Response:
     """Return the last export timestamp."""
@@ -109,7 +99,7 @@ def export_chats() -> tuple[Response, int] | Response:
             workspace_path=workspace_path,
             exclusion_rules=exclusion_rules(),
             since=since,
-            last_export_ms=_read_last_export_ms(since),
+            last_export_ms=read_last_export_ms(since, state=_get_export_state()),
             out_dir="",
             include_composer=True,
             include_cli=False,

models/__init__.py

@@ -3,7 +3,7 @@
 from models.conversation import Bubble, Composer, Conversation, WorkspaceLocalComposer
 from models.errors import SchemaError
 from models.parse_warnings import ParseWarningCollector
-from models.export import ExportEntry
+from models.export import CollectedExportEntry, ExportEntry
 from models.search import ConversationSummary, SearchResult
 from models.workspace import Workspace
 
@@ -16,6 +16,7 @@
     "Composer",
     "Conversation",
     "ConversationSummary",
+    "CollectedExportEntry",
     "ExportEntry",
     "ParseWarningCollector",
     "SchemaError",

models/export.py

@@ -1,11 +1,23 @@
 from __future__ import annotations
 
 from dataclasses import dataclass, field
-from typing import Any
+from typing import Any, TypedDict
 
 from models.from_dict_validation import require_dict, require_non_empty_str_fields
 
 
+class CollectedExportEntry(TypedDict):
+    """One exportable conversation with rendered markdown (engine/CLI collection)."""
+
+    id: str
+    rel_path: str
+    content: str
+    out_path: str
+    updatedAt: int
+    title: str
+    workspace: str
+
+
 @dataclass(frozen=True)
 class ExportEntry:
     """One line of manifest.jsonl; log_id / title / workspace required, timestamps optional."""

scripts/export.py

@@ -31,12 +31,11 @@
         sys.path.insert(0, str(_project_root))
 
 from models import ExportEntry, SchemaError  # noqa: E402
-from services.export_engine import collect_export_entries  # noqa: E402
+from services.export_engine import collect_export_entries, read_last_export_ms  # noqa: E402
 from utils.exclusion_rules import (  # noqa: E402
     load_rules,
     resolve_exclusion_rules_path,
 )
-from utils.path_helpers import to_epoch_ms  # noqa: E402
 from utils.workspace_path import resolve_workspace_path  # noqa: E402
 
 _logger = logging.getLogger(__name__)
@@ -168,23 +167,6 @@ def parse_args() -> ExportCliOptions:
     }
 
 
-def _read_last_export_ms(state_path: str, since: Literal["all", "last"]) -> int:
-    if since != "last" or not os.path.isfile(state_path):
-        return 0
-    try:
-        with open(state_path, "r", encoding="utf-8") as f:
-            st = json.load(f)
-        ts = st.get("lastExportTime")
-        if ts:
-            return to_epoch_ms(ts)
-    except (json.JSONDecodeError, ValueError, OSError) as e:
-        _logger.warning(
-            "Could not read last export timestamp; defaulting to full export: %s",
-            e,
-        )
-    return 0
-
-
 def main() -> None:
     configure_cli_logging()
     opts = parse_args()
@@ -198,7 +180,7 @@ def main() -> None:
 
     state_dir = get_global_state_dir()
     state_path = os.path.join(state_dir, "export_state.json")
-    last_export = _read_last_export_ms(state_path, since)
+    last_export = read_last_export_ms(since, state_path=state_path)
 
     exported = collect_export_entries(
         workspace_path=workspace_path,

services/export_engine.py

@@ -8,9 +8,10 @@
 import sqlite3
 from dataclasses import dataclass
 from datetime import datetime
-from typing import Any, Literal, TypedDict
+from typing import Any, Literal
 
 from models import Bubble
+from models.export import CollectedExportEntry
 from services.summary_cache import fingerprint_workspace_storage, nocache_enabled
 from services.workspace_context import (
     WorkspaceContext,
@@ -49,16 +50,32 @@
 SinceMode = Literal["all", "last"]
 
 
-class ExportEntry(TypedDict):
-    """One exportable conversation with rendered markdown."""
-
-    id: str
-    rel_path: str
-    content: str
-    out_path: str
-    updatedAt: int
-    title: str
-    workspace: str
+def read_last_export_ms(
+    since: SinceMode,
+    *,
+    state_path: str | None = None,
+    state: dict[str, Any] | None = None,
+) -> int:
+    """Return last-export epoch ms for ``since=last``; 0 for a full export."""
+    if since != "last":
+        return 0
+    ts: Any = None
+    if state is not None:
+        ts = state.get("lastExportTime")
+    elif state_path is not None and os.path.isfile(state_path):
+        try:
+            with open(state_path, "r", encoding="utf-8") as f:
+                st = json.load(f)
+            if isinstance(st, dict):
+                ts = st.get("lastExportTime")
+        except (json.JSONDecodeError, ValueError, OSError) as e:
+            _logger.warning(
+                "Could not read last export timestamp; defaulting to full export: %s",
+                e,
+            )
+    if ts:
+        return to_epoch_ms(ts)
+    return 0
 
 
 @dataclass(frozen=True)
@@ -209,9 +226,9 @@ def _collect_ide_export_entries(
     last_export_ms: int,
     today: str,
     out_dir: str,
-) -> list[ExportEntry]:
+) -> list[CollectedExportEntry]:
     ctx = orch.ctx
-    exported: list[ExportEntry] = []
+    exported: list[CollectedExportEntry] = []
     for row in db_data.ide_composer_rows:
         composer_id = row["key"].split(":")[1]
         try:
@@ -239,6 +256,8 @@ def _collect_ide_export_entries(
         updated_at = to_epoch_ms(cd.get("lastUpdatedAt")) or to_epoch_ms(
             cd.get("createdAt"),
         )
+        # Intentional behavior change vs legacy CLI: fall back to createdAt when
+        # lastUpdatedAt is absent (affects timestamps, filenames, and --since last).
         if since == "last" and updated_at <= last_export_ms:
             continue
 
@@ -325,7 +344,7 @@ def _collect_ide_export_entries(
             workspace_info={"ws_slug": ws_slug, "ws_display_name": ws_display_name},
         )
 
-        rel_path = os.path.join(today, ws_slug, "chat", filename)
+        rel_path = os.path.relpath(out_path, out_dir)
         exported.append({
             "id": composer_id,
             "rel_path": rel_path,
@@ -345,8 +364,8 @@ def _collect_cli_export_entries(
     last_export_ms: int,
     today: str,
     out_dir: str,
-) -> list[ExportEntry]:
-    exported: list[ExportEntry] = []
+) -> list[CollectedExportEntry]:
+    exported: list[CollectedExportEntry] = []
     try:
         cli_projects = list_cli_projects(get_cli_chats_path())
     except Exception as e:  # noqa: BLE001 — log and skip CLI enumeration on any failure
@@ -446,7 +465,7 @@ def _collect_cli_export_entries(
                 bubbles=bubbles,
                 title_override=title,
             )
-            rel_path = os.path.join(today, ws_slug_cli, "cli", filename)
+            rel_path = os.path.relpath(out_path, out_dir)
             exported.append({
                 "id": session_id,
                 "rel_path": rel_path,
@@ -469,14 +488,14 @@ def collect_export_entries(
     include_composer: bool = True,
     include_cli: bool = True,
     nocache: bool = False,
-) -> list[ExportEntry]:
+) -> list[CollectedExportEntry]:
     """Collect exportable conversations (IDE + CLI) via shared orchestration."""
     effective_nocache = nocache_enabled(request_nocache=nocache)
     orch = prepare_workspace_orchestration(
         workspace_path, exclusion_rules, nocache=effective_nocache,
     )
     today = datetime.now().strftime("%Y-%m-%d")
-    exported: list[ExportEntry] = []
+    exported: list[CollectedExportEntry] = []
 
     if include_composer:
         db_data = load_global_db_export_data(orch)