chore(publication): close pre-integration release blockers

scripts/add_hwfit_models.py

@@ -1,7 +1,7 @@
 #!/usr/bin/env python3
 """
 add_hwfit_models.py — bulk-add Hugging Face models to the hwfit catalog
-(services/hwfit/data/hf_models.json).
+(DATA_DIR/hwfit/hf_models.json, mutable user data).
 
 Adds:
   * every model from one or more HF authors (e.g. cyankiwi's AWQ quants)
@@ -28,8 +28,10 @@ from datetime import datetime
 from huggingface_hub import HfApi, hf_hub_download
 from huggingface_hub.utils import EntryNotFoundError, RepositoryNotFoundError
 
-DATA_PATH = os.path.join(os.path.dirname(__file__), "..", "services", "hwfit", "data", "hf_models.json")
-DATA_PATH = os.path.abspath(DATA_PATH)
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from services.hwfit.models import model_catalog_path
+
+DATA_PATH = model_catalog_path()
 
 # Official / major model-provider orgs to refresh into the Cookbook catalog.
 # Keep this broad enough that new first-party releases appear after running the
@@ -437,8 +439,12 @@ def _is_likely_catalog_model(mi):
 
 
 def main():
-    with open(DATA_PATH, encoding="utf-8") as f:
-        catalog = json.load(f)
+    os.makedirs(os.path.dirname(DATA_PATH), exist_ok=True)
+    if os.path.exists(DATA_PATH):
+        with open(DATA_PATH, encoding="utf-8") as f:
+            catalog = json.load(f)
+    else:
+        catalog = []
     by_name = {m["name"]: m for m in catalog}
     existing = set(by_name)
 

scripts/backfill_model_release_dates.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Backfill release_date on entries in services/hwfit/data/hf_models.json.
+"""Backfill release_date on entries in DATA_DIR/hwfit/hf_models.json.
 
 Why: the `newest` sort in the cookbook ranks rows by release_date. Anything
 missing a date sorts to the bottom. This script pulls `created_at` from the
@@ -30,7 +30,10 @@ except ImportError:
     sys.exit(1)
 
 
-CATALOG_PATH = Path(__file__).resolve().parent.parent / "services" / "hwfit" / "data" / "hf_models.json"
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from services.hwfit.models import model_catalog_path
+
+CATALOG_PATH = Path(model_catalog_path())
 
 
 def fetch_release_date(api: HfApi, repo_id: str) -> str | None:

scripts/import_from_vllm_recipes.py

@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 """Import models from the upstream vllm-project/recipes catalog into our
-local hf_models.json. Two modes:
+runtime DATA_DIR/hwfit/hf_models.json. Two modes:
 
   --update-existing  Stamp min_vllm_version + vllm_recipe=True on rows we
                      already carry. Cheap, no HF API calls.
@@ -44,7 +44,10 @@ except ImportError:
     HfHubHTTPError = Exception
 
 
-CATALOG_PATH = Path(__file__).resolve().parent.parent / "services" / "hwfit" / "data" / "hf_models.json"
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+from services.hwfit.models import model_catalog_path
+
+CATALOG_PATH = Path(model_catalog_path())
 RECIPES_TREE_URL = (
     "https://api.github.com/repos/vllm-project/recipes/git/trees/main?recursive=1"
 )
@@ -253,8 +256,12 @@ def main():
     if not args.update_existing and not args.add_missing:
         args.update_existing = args.add_missing = True
 
-    with CATALOG_PATH.open(encoding="utf-8") as f:
-        catalog = json.load(f)
+    CATALOG_PATH.parent.mkdir(parents=True, exist_ok=True)
+    if CATALOG_PATH.exists():
+        with CATALOG_PATH.open(encoding="utf-8") as f:
+            catalog = json.load(f)
+    else:
+        catalog = []
     by_name = {m.get("name"): m for m in catalog if m.get("name")}
 
     client = httpx.Client(follow_redirects=True)