refactor(services): centralize workspace determination in resolve_workspace_context (#91)

api/export_api.py

@@ -22,17 +22,13 @@
 from utils.text_extract import extract_text_from_bubble, slug
 from utils.exclusion_rules import build_searchable_text, is_excluded_by_rules
 from utils.cursor_md_exporter import cursor_ide_chat_to_markdown
+from services.workspace_context import resolve_workspace_context_minimal
 from services.workspace_db import (
-    build_composer_id_to_workspace_id,
-    collect_workspace_entries,
     load_bubble_map,
     load_code_block_diff_map,
     open_global_db,
 )
-from services.workspace_resolver import (
-    create_project_name_to_workspace_id_map,
-    lookup_workspace_display_name,
-)
+from services.workspace_resolver import lookup_workspace_display_name
 
 bp = Blueprint("export_api", __name__)
 _logger = logging.getLogger(__name__)
@@ -102,9 +98,9 @@ def export_chats():
                 last_export_ms = to_epoch_ms(ts_str)
 
         # ── Workspace scanning via service layer ──────────────────────────────
-        workspace_entries = collect_workspace_entries(workspace_path)
-        composer_id_to_ws = build_composer_id_to_workspace_id(workspace_path, workspace_entries)
-        project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
+        ctx = resolve_workspace_context_minimal(workspace_path)
+        workspace_entries = ctx.workspace_entries
+        composer_id_to_ws = ctx.composer_id_to_workspace_id
 
         # Build display-name and slug maps
         ws_id_to_slug: dict[str, str] = {}

scripts/export.py

@@ -53,18 +53,15 @@
     cursor_ide_chat_to_markdown,
 )
 from models import ExportEntry, SchemaError  # noqa: E402
+from services.workspace_context import (  # noqa: E402
+    enrich_workspace_context_from_global_db,
+    resolve_workspace_context,
+)
 from services.workspace_db import (  # noqa: E402
-    build_composer_id_to_workspace_id,
-    collect_invalid_workspace_ids,
-    collect_workspace_entries,
-    load_bubble_map,
     load_code_block_diff_map,
-    load_project_layouts_map,
     open_global_db,
 )
 from services.workspace_resolver import (  # noqa: E402
-    create_project_name_to_workspace_id_map,
-    create_workspace_path_to_id_map,
     determine_project_for_conversation,
     infer_invalid_workspace_aliases,
     lookup_workspace_display_name,
@@ -203,11 +200,12 @@ def main():
             )
 
     # ── Workspace scanning via service layer ──────────────────────────────────
-    workspace_entries = collect_workspace_entries(workspace_path)
-    invalid_workspace_ids = collect_invalid_workspace_ids(workspace_entries)
-    project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
-    workspace_path_map = create_workspace_path_to_id_map(workspace_entries)
-    composer_id_to_ws = build_composer_id_to_workspace_id(workspace_path, workspace_entries)
+    ctx = resolve_workspace_context(workspace_path)
+    workspace_entries = ctx.workspace_entries
+    invalid_workspace_ids = ctx.invalid_workspace_ids
+    project_name_map = ctx.project_name_to_workspace_id
+    workspace_path_map = ctx.workspace_path_to_id
+    composer_id_to_ws = ctx.composer_id_to_workspace_id
 
     # Build display-name and slug maps from workspace entries.
     # Entries whose workspace.json cannot be resolved are omitted so the
@@ -235,8 +233,14 @@ def main():
                 global_db_path,
             )
         else:
-            project_layouts_map = load_project_layouts_map(global_db)
-            bubble_map = load_bubble_map(global_db)
+            ctx = enrich_workspace_context_from_global_db(
+                ctx,
+                global_db,
+                populate_project_layouts=True,
+                populate_bubble_map=True,
+            )
+            project_layouts_map = ctx.project_layouts_map
+            bubble_map = ctx.bubble_map
             code_block_diff_map = load_code_block_diff_map(global_db)
 
             try:

services/workspace_context.py

@@ -0,0 +1,135 @@
+"""Workspace determination ceremony — single orchestrator for shared maps."""
+
+from __future__ import annotations
+
+import sqlite3
+from dataclasses import dataclass, replace
+
+from services.workspace_db import (
+    build_composer_id_to_workspace_id,
+    build_composer_id_to_workspace_id_cached,
+    collect_invalid_workspace_ids,
+    collect_workspace_entries,
+    load_bubble_map,
+    load_project_layouts_map,
+)
+from services.workspace_resolver import (
+    create_project_name_to_workspace_id_map,
+    create_workspace_path_to_id_map,
+)
+
+
+@dataclass(frozen=True)
+class WorkspaceContext:
+    """Precomputed workspace-resolution maps for conversation assignment."""
+
+    workspace_entries: list[dict]
+    invalid_workspace_ids: set[str]
+    composer_id_to_workspace_id: dict[str, str]
+    project_name_to_workspace_id: dict[str, str]
+    workspace_path_to_id: dict[str, str]
+    project_layouts_map: dict[str, list]
+    bubble_map: dict[str, dict]
+
+
+def _entries(
+    workspace_path: str,
+    workspace_entries: list[dict] | None,
+) -> list[dict]:
+    if workspace_entries is not None:
+        return workspace_entries
+    return collect_workspace_entries(workspace_path)
+
+
+def _assemble_context(
+    entries: list[dict],
+    *,
+    invalid_workspace_ids: set[str],
+    workspace_path_to_id: dict[str, str],
+    composer_id_to_workspace_id: dict[str, str],
+) -> WorkspaceContext:
+    return WorkspaceContext(
+        workspace_entries=entries,
+        invalid_workspace_ids=invalid_workspace_ids,
+        composer_id_to_workspace_id=composer_id_to_workspace_id,
+        project_name_to_workspace_id=create_project_name_to_workspace_id_map(entries),
+        workspace_path_to_id=workspace_path_to_id,
+        project_layouts_map={},
+        bubble_map={},
+    )
+
+
+def resolve_workspace_context(
+    workspace_path: str,
+    *,
+    workspace_entries: list[dict] | None = None,
+) -> WorkspaceContext:
+    """Full workspace maps with an uncached composer→workspace scan (CLI export)."""
+    entries = _entries(workspace_path, workspace_entries)
+    return _assemble_context(
+        entries,
+        invalid_workspace_ids=collect_invalid_workspace_ids(entries),
+        workspace_path_to_id=create_workspace_path_to_id_map(entries),
+        composer_id_to_workspace_id=build_composer_id_to_workspace_id(
+            workspace_path, entries,
+        ),
+    )
+
+
+def resolve_workspace_context_cached(
+    workspace_path: str,
+    rules: list,
+    *,
+    workspace_entries: list[dict] | None = None,
+    nocache: bool = False,
+) -> WorkspaceContext:
+    """Full workspace maps with a mtime-keyed composer map (listing / tabs)."""
+    entries = _entries(workspace_path, workspace_entries)
+    return _assemble_context(
+        entries,
+        invalid_workspace_ids=collect_invalid_workspace_ids(entries),
+        workspace_path_to_id=create_workspace_path_to_id_map(entries),
+        composer_id_to_workspace_id=build_composer_id_to_workspace_id_cached(
+            workspace_path, entries, rules, nocache=nocache,
+        ),
+    )
+
+
+def resolve_workspace_context_minimal(
+    workspace_path: str,
+    *,
+    workspace_entries: list[dict] | None = None,
+) -> WorkspaceContext:
+    """Entries, project-name, and composer maps only (HTTP export).
+
+    Args:
+        workspace_path: Cursor ``workspaceStorage`` root.
+        workspace_entries: Pre-collected entries; when ``None``, scanned from disk.
+    """
+    entries = _entries(workspace_path, workspace_entries)
+    return _assemble_context(
+        entries,
+        invalid_workspace_ids=set(),
+        workspace_path_to_id={},
+        composer_id_to_workspace_id=build_composer_id_to_workspace_id(
+            workspace_path, entries,
+        ),
+    )
+
+
+def enrich_workspace_context_from_global_db(
+    ctx: WorkspaceContext,
+    global_db: sqlite3.Connection,
+    *,
+    populate_project_layouts: bool = False,
+    populate_bubble_map: bool = False,
+) -> WorkspaceContext:
+    """Return *ctx* with global KV maps loaded from an open global DB connection."""
+    updates: dict = {}
+    if populate_project_layouts:
+        updates["project_layouts_map"] = load_project_layouts_map(global_db)
+    if populate_bubble_map:
+        updates["bubble_map"] = load_bubble_map(global_db)
+    if not updates:
+        return ctx
+    return replace(ctx, **updates)

services/workspace_listing.py

@@ -24,10 +24,9 @@
     nocache_enabled,
     set_cached_projects,
 )
+from services.workspace_context import resolve_workspace_context_cached
 from services.workspace_db import (
     COMPOSER_ROWS_WITH_HEADERS_SQL,
-    build_composer_id_to_workspace_id_cached,
-    collect_invalid_workspace_ids,
     collect_workspace_entries,
     global_storage_db_path,
     load_project_layouts_for_composer,
@@ -36,8 +35,6 @@
 )
 from utils.workspace_path import get_cli_chats_path
 from services.workspace_resolver import (
-    create_project_name_to_workspace_id_map,
-    create_workspace_path_to_id_map,
     determine_project_for_conversation,
     infer_invalid_workspace_aliases,
     infer_workspace_name_from_context,
@@ -125,13 +122,16 @@ def _build_workspace_projects_uncached(
     nocache: bool,
 ) -> tuple[list[dict], list[dict]]:
     parse_warnings = ParseWarningCollector()
-    invalid_workspace_ids = collect_invalid_workspace_ids(workspace_entries)
-
-    project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
-    workspace_path_map = create_workspace_path_to_id_map(workspace_entries)
-    composer_id_to_ws = build_composer_id_to_workspace_id_cached(
-        workspace_path, workspace_entries, rules, nocache=nocache,
+    ctx = resolve_workspace_context_cached(
+        workspace_path,
+        rules,
+        workspace_entries=workspace_entries,
+        nocache=nocache,
     )
+    invalid_workspace_ids = ctx.invalid_workspace_ids
+    project_name_map = ctx.project_name_to_workspace_id
+    workspace_path_map = ctx.workspace_path_to_id
+    composer_id_to_ws = ctx.composer_id_to_workspace_id
 
     conversation_map: dict[str, list] = {}
 

services/workspace_tabs.py

@@ -27,10 +27,9 @@
     nocache_enabled,
     set_cached_tab_summaries,
 )
+from services.workspace_context import resolve_workspace_context_cached
 from services.workspace_db import (
     COMPOSER_ROWS_WITH_HEADERS_SQL,
-    build_composer_id_to_workspace_id_cached,
-    collect_invalid_workspace_ids,
     collect_workspace_entries,
     global_storage_db_path,
     load_bubbles_for_composer,
@@ -43,8 +42,6 @@
 )
 from utils.workspace_path import get_cli_chats_path
 from services.workspace_resolver import (
-    create_project_name_to_workspace_id_map,
-    create_workspace_path_to_id_map,
     determine_project_for_conversation,
     infer_invalid_workspace_aliases,
     lookup_workspace_display_name,
@@ -489,12 +486,16 @@ def _build_workspace_tab_summaries_uncached(
     parse_warnings = ParseWarningCollector()
     response: dict = {"tabs": []}
 
-    invalid_workspace_ids = collect_invalid_workspace_ids(workspace_entries)
-    project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
-    workspace_path_map = create_workspace_path_to_id_map(workspace_entries)
-    composer_id_to_ws = build_composer_id_to_workspace_id_cached(
-        workspace_path, workspace_entries, rules, nocache=nocache,
+    ctx = resolve_workspace_context_cached(
+        workspace_path,
+        rules,
+        workspace_entries=workspace_entries,
+        nocache=nocache,
     )
+    invalid_workspace_ids = ctx.invalid_workspace_ids
+    project_name_map = ctx.project_name_to_workspace_id
+    workspace_path_map = ctx.workspace_path_to_id
+    composer_id_to_ws = ctx.composer_id_to_workspace_id
     matching_ws_ids = _build_matching_ws_ids(workspace_id, workspace_path, workspace_entries)
 
     with open_global_db(workspace_path) as (global_db, _):
@@ -635,13 +636,12 @@ def assemble_single_tab(
     """
     parse_warnings = ParseWarningCollector()
 
-    workspace_entries = collect_workspace_entries(workspace_path)
-    invalid_workspace_ids = collect_invalid_workspace_ids(workspace_entries)
-    project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
-    workspace_path_map = create_workspace_path_to_id_map(workspace_entries)
-    composer_id_to_ws = build_composer_id_to_workspace_id_cached(
-        workspace_path, workspace_entries, rules,
-    )
+    ctx = resolve_workspace_context_cached(workspace_path, rules)
+    workspace_entries = ctx.workspace_entries
+    invalid_workspace_ids = ctx.invalid_workspace_ids
+    project_name_map = ctx.project_name_to_workspace_id
+    workspace_path_map = ctx.workspace_path_to_id
+    composer_id_to_ws = ctx.composer_id_to_workspace_id
     matching_ws_ids = _build_matching_ws_ids(workspace_id, workspace_path, workspace_entries)
 
     with open_global_db(workspace_path) as (global_db, _):
@@ -768,13 +768,12 @@ def assemble_workspace_tabs(
     parse_warnings = ParseWarningCollector()
     response: dict = {"tabs": []}
 
-    workspace_entries = collect_workspace_entries(workspace_path)
-    invalid_workspace_ids = collect_invalid_workspace_ids(workspace_entries)
-    project_name_map = create_project_name_to_workspace_id_map(workspace_entries)
-    workspace_path_map = create_workspace_path_to_id_map(workspace_entries)
-    composer_id_to_ws = build_composer_id_to_workspace_id_cached(
-        workspace_path, workspace_entries, rules,
-    )
+    ctx = resolve_workspace_context_cached(workspace_path, rules)
+    workspace_entries = ctx.workspace_entries
+    invalid_workspace_ids = ctx.invalid_workspace_ids
+    project_name_map = ctx.project_name_to_workspace_id
+    workspace_path_map = ctx.workspace_path_to_id
+    composer_id_to_ws = ctx.composer_id_to_workspace_id
     matching_ws_ids = _build_matching_ws_ids(workspace_id, workspace_path, workspace_entries)
 
     bubble_map: dict[str, dict] = {}
@@ -840,20 +839,20 @@ def _safe_fetchall(query: str, params: tuple = ()) -> list:
             if len(parts) < 2:
                 continue
             chat_id = parts[1]
-            ctx = _loads_kv_value_logged(row["key"], row["value"])
-            if not isinstance(ctx, dict):
+            mrc = _loads_kv_value_logged(row["key"], row["value"])
+            if not isinstance(mrc, dict):
                 continue
 
             # Per-bubble context map (needs the contextId at parts[2])
             if len(parts) >= 3:
                 context_id = parts[2]
                 message_request_context_map.setdefault(chat_id, []).append({
-                    **ctx,
+                    **mrc,
                     "contextId": context_id,
                 })
 
             # Project-layout map (root paths used by the resolver)
-            layouts = ctx.get("projectLayouts")
+            layouts = mrc.get("projectLayouts")
             if isinstance(layouts, list):
                 project_layouts_map.setdefault(chat_id, [])
                 for layout in layouts: