Revision defaults to empty string

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "truss"
-version = "0.11.26"
+version = "0.11.27rc500"
 description = "A seamless bridge from model development to model delivery"
 authors = [
     { name = "Pankaj Gupta", email = "no-reply@baseten.co" },

truss-transfer/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "truss_transfer"
-version = "0.0.38"
+version = "0.0.39"
 edition = "2021"
 
 [lib]

truss-transfer/src/bindings.rs

@@ -114,8 +114,8 @@ impl PyModelRepo {
     #[new]
     #[pyo3(signature = (
         repo_id,
-        revision,
-        volume_folder,
+        revision = "".to_string(),
+        volume_folder = "".to_string(),
         kind = "hf".to_string(),
         allow_patterns = None,
         ignore_patterns = None,

truss-transfer/src/create/basetenpointer.rs

@@ -72,7 +72,7 @@ mod tests {
             },
             ModelRepo {
                 repo_id: "gs://test-bucket/model-path".to_string(),
-                revision: "main".to_string(), // Ignored for GCS
+                revision: "".to_string(), // Not needed for GCS
                 allow_patterns: Some(vec!["*.safetensors".to_string()]),
                 ignore_patterns: Some(vec!["*.md".to_string()]),
                 kind: ResolutionType::Gcs,
@@ -249,7 +249,7 @@ mod tests {
         // Test Azure support with a mock repository
         let model_repos = vec![ModelRepo {
             repo_id: "azure://testaccount/testcontainer/model.bin".to_string(),
-            revision: "main".to_string(),
+            revision: "".to_string(), // Not needed for Azure
             runtime_secret_name: "azure-storage".to_string(),
             volume_folder: "test_azure_model".to_string(),
             kind: ResolutionType::Azure,

truss/base/truss_config.py

@@ -141,7 +141,7 @@ class ModelRepoSourceKind(str, enum.Enum):
 
 class ModelRepo(custom_types.ConfigModel):
     repo_id: Annotated[str, pydantic.StringConstraints(min_length=1)]
-    revision: Optional[Annotated[str, pydantic.StringConstraints(min_length=1)]] = None
+    revision: str = ""
     allow_patterns: Optional[list[str]] = None
     ignore_patterns: Optional[list[str]] = None
     volume_folder: Optional[
@@ -151,6 +151,13 @@ class ModelRepo(custom_types.ConfigModel):
     kind: ModelRepoSourceKind = ModelRepoSourceKind.HF
     runtime_secret_name: str = "hf_access_token"
 
+    @pydantic.field_validator("revision")
+    @classmethod
+    def _validate_revision(cls, v: str) -> str:
+        if len(v) == 1:
+            raise ValueError("revision must be empty or at least 2 characters")
+        return v
+
     @property
     def runtime_path(self) -> pathlib.Path:
         assert self.volume_folder is not None
@@ -161,11 +168,14 @@ def _check_v2_requirements(cls, v) -> str:
         use_volume = v.get("use_volume", False)
         if not use_volume:
             return v
-        if v.get("kind") == ModelRepoSourceKind.HF.value and v.get("revision") is None:
+        revision = v.get("revision") or ""
+        kind = v.get("kind")
+        is_hf = kind is None or kind == ModelRepoSourceKind.HF.value
+        if is_hf and not revision:
             logger.warning(
-                "the key `revision: str` is required for use_volume=True huggingface repos. For S3/GCS/Azure repos, set it to any non-empty string."
+                "the key `revision: str` is required for use_volume=True huggingface repos."
             )
-            raise_insufficent_revision(v.get("repo_id"), v.get("revision"))
+            raise_insufficent_revision(v.get("repo_id"), revision)
         if v.get("volume_folder") is None or len(v["volume_folder"]) == 0:
             raise ValueError(
                 "the key `volume_folder: str` is required for `use_volume=True` repos."

truss/templates/cache.Dockerfile.jinja

@@ -20,6 +20,6 @@ COPY ./{{credential}} ${APP_HOME}/{{credential}}
 
 {% for repo, hf_dir in models.items() %}
         {% for file in hf_dir.files %}
-{{ "RUN --mount=type=secret,id=" + hf_access_token_file_name + ",dst=/etc/secrets/" + hf_access_token_file_name if use_hf_secret else "RUN" }} python3 /cache_warmer.py {{file}} {{repo}} {% if hf_dir.revision != None %}{{hf_dir.revision}}{% endif %}
+{{ "RUN --mount=type=secret,id=" + hf_access_token_file_name + ",dst=/etc/secrets/" + hf_access_token_file_name if use_hf_secret else "RUN" }} python3 /cache_warmer.py {{file}} {{repo}} {% if hf_dir.revision %}{{hf_dir.revision}}{% endif %}
         {% endfor %}
 {% endfor %}

truss/tests/test_config.py

@@ -317,7 +317,7 @@ def test_huggingface_cache_single_model_default_revision(default_config):
     new_config["model_cache"] = [{"repo_id": "test/model", "use_volume": False}]
 
     assert new_config == config.to_dict(verbose=False)
-    assert config.to_dict(verbose=True)["model_cache"][0].get("revision") is None
+    assert config.to_dict(verbose=True)["model_cache"][0].get("revision") == ""
 
 
 def test_huggingface_cache_single_model_non_default_revision_v1():
@@ -354,7 +354,7 @@ def test_huggingface_cache_multiple_models_default_revision(default_config):
         "model_cache"
     ]
     assert config.to_dict(verbose=True)["model_cache"][0].get("revision") == "main"
-    assert config.to_dict(verbose=True)["model_cache"][1].get("revision") is None
+    assert config.to_dict(verbose=True)["model_cache"][1].get("revision") == ""
 
 
 def test_huggingface_cache_multiple_models_mixed_revision(default_config):
@@ -377,7 +377,7 @@ def test_huggingface_cache_multiple_models_mixed_revision(default_config):
     ]
 
     assert new_config == config.to_dict(verbose=False)
-    assert config.to_dict(verbose=True)["model_cache"][0].get("revision") is None
+    assert config.to_dict(verbose=True)["model_cache"][0].get("revision") == ""
     assert config.to_dict(verbose=True)["model_cache"][1].get("revision") == "not-main2"