Initial check-in of changes to llms() API improving model name resolution

.github/workflows/build-binaries.yml

@@ -263,7 +263,7 @@ jobs:
           githubToken: ${{ github.token }}
           install: |
             apt-get update
-            apt-get install -y --no-install-recommends python3 python3-pip python3-venv cargo
+            apt-get install -y --no-install-recommends python3 python3-pip python3-venv python3-dev cargo
             pip3 install -U pip
           # Create and use a virtual environment to avoid the externally-managed-environment error
           run: |
@@ -307,7 +307,7 @@ jobs:
           image: alpine:latest
           options: -v ${{ github.workspace }}:/io -w /io
           run: |
-            apk add python3 py3-pip rust
+            apk add python3 python3-dev py3-pip rust
             python -m venv .venv
             .venv/bin/pip3 install dist/${{ env.PACKAGE_NAME }}-*.whl --force-reinstall
             .venv/bin/${{ env.EXECUTABLE_NAME }} --help
@@ -348,7 +348,7 @@ jobs:
           distro: alpine_latest
           githubToken: ${{ github.token }}
           install: |
-            apk add python3 py3-pip rust
+            apk add python3 python3-dev py3-pip rust
           run: |
             python -m venv .venv
             .venv/bin/pip3 install dist/${{ env.PACKAGE_NAME }}-*.whl --force-reinstall

.gitignore

@@ -3,6 +3,10 @@
 *.pyc
 __pycache__
 
+
+# may contain sensitive data
+pytest.ini
+
 # 
 # Artifacts from the Rust client generation process
 #

.vscode/settings.json

@@ -0,0 +1,7 @@
+{
+    "python.testing.pytestArgs": [
+        "tests"
+    ],
+    "python.testing.unittestEnabled": false,
+    "python.testing.pytestEnabled": true
+}

README.md

@@ -189,3 +189,7 @@ uv run pytest tests
 
 If you need to get the latest OpenAPI SDK, you can run
 `./scripts/generate-python-api-client.sh`.
+
+## Testing
+We use pytest to run tests. Copy `pytest.ini.template` to `pytest.ini` and 
+replace the values of environment variables 

pyproject.toml

@@ -40,11 +40,14 @@ classifiers = [
 dependencies = [
   "attrs==24.2.0",
   "httpx==0.28.1",
+  "huggingface-hub>=0.34.3",
+  "ollama>=0.4.7",
+  "pyiceberg==0.9.0",
   "python-dateutil==2.9.0.post0",
 ]
 
 [project.optional-dependencies]
-ai = ["huggingface-hub==0.30.2", "ollama==0.4.7"]
+ai = ["huggingface-hub==0.34.3", "ollama==0.4.7"]
 iceberg = ["polars==1.27.1", "pyarrow==19.0.1", "pyiceberg==0.9.0"]
 all = ["tower[ai,iceberg]"]
 
@@ -65,5 +68,6 @@ dev = [
   "openapi-python-client==0.24.3",
   "pytest==8.3.5",
   "pytest-httpx==0.35.0",
+  "pytest-env>=1.1.3",
   "pyiceberg[sql-sqlite]==0.9.0",
 ]

pytest.ini.template

@@ -0,0 +1,12 @@
+[pytest]
+pythonpath = src
+testpaths = tests
+python_files = test_*.py
+python_classes = Test*
+python_functions = test_*
+addopts = -v --tb=short
+filterwarnings =
+    ignore::DeprecationWarning
+    ignore::PendingDeprecationWarning
+env =
+    TOWER_INFERENCE_ROUTER_API_KEY=<hf_1234567890>

src/tower/_context.py

@@ -1,13 +1,16 @@
 import os
 
 class TowerContext:
-    def __init__(self, tower_url: str, environment: str, api_key: str = None, hugging_face_provider: str = None, hugging_face_api_key: str = None, jwt: str = None):
+    def __init__(self, tower_url: str, environment: str, api_key: str = None, 
+                 inference_router: str = None, inference_router_api_key: str = None,
+                 inference_provider: str = None, jwt: str = None):
         self.tower_url = tower_url
         self.environment = environment
         self.api_key = api_key
         self.jwt = jwt
-        self.hugging_face_provider = hugging_face_provider
-        self.hugging_face_api_key = hugging_face_api_key
+        self.inference_router = inference_router
+        self.inference_router_api_key = inference_router_api_key
+        self.inference_provider = inference_provider
 
     def is_local(self) -> bool:
         if self.environment is None or self.environment == "":
@@ -24,17 +27,18 @@ def build(cls):
         tower_api_key = os.getenv("TOWER_API_KEY")
         tower_jwt = os.getenv("TOWER_JWT")
 
-        # NOTE: These are experimental, used only for our experimental Hugging
-        # Face integration for LLMs.
-        hugging_face_provider = os.getenv("TOWER_HUGGING_FACE_PROVIDER")
-        hugging_face_api_key = os.getenv("TOWER_HUGGING_FACE_API_KEY")
+        # Replaces the deprecated hugging_face_provider and hugging_face_api_key
+        inference_router = os.getenv("TOWER_INFERENCE_ROUTER")
+        inference_router_api_key = os.getenv("TOWER_INFERENCE_ROUTER_API_KEY")
+        inference_provider = os.getenv("TOWER_INFERENCE_PROVIDER")
 
         return cls(
             tower_url = tower_url,
             environment = tower_environment,
             api_key = tower_api_key,
-            hugging_face_provider = hugging_face_provider,
-            hugging_face_api_key = hugging_face_api_key,
+            inference_router = inference_router,
+            inference_router_api_key = inference_router_api_key,
+            inference_provider = inference_provider,
             jwt = tower_jwt,
         )