chore: cleanup todo comments throughout the codebase (#375)

.github/workflows/e2e.yaml

@@ -44,7 +44,6 @@ jobs:
           with:
             python-version-file: 'pyproject.toml'
 
-        # TODO @JPERRY: We should be generating the proto files before this...
         - name: Install Python Deps
           run: python -m pip install ".[dev,e2e-test]"
 

.github/workflows/pytest.yaml

@@ -31,7 +31,6 @@ jobs:
         with:
           python-version-file: 'pyproject.toml'
 
-      # TODO @JPERRY: This repeater image is floating in the void. Give it a home in this repo???
       - name: Run Repeater
         run: docker run -p 50051:50051 -d ghcr.io/defenseunicorns/leapfrogai/repeater:0.3.3
 

Makefile

@@ -30,7 +30,6 @@ gen-python: ## Generate the protobufs for the OpenAI typing within the leapfroga
 			--grpc_python_out=src/. \
 			src/leapfrogai_api/types/proto/leapfrogai_api/types/**/*.proto
 
-# TODO: Make the port a variable with an uncommon default
 local-registry: ## Start up a local container registry. Errors in this target are ignored.
 	-docker run -d -p 5000:5000 --restart=always --name registry registry:2
 

packages/llama-cpp-python/Dockerfile

@@ -10,7 +10,6 @@ ARG REPO_ID=TheBloke/SynthIA-7B-v2.0-GGUF
 ARG FILENAME=synthia-7b-v2.0.Q4_K_M.gguf
 ARG REVISION=3f65d882253d1f15a113dabf473a7c02a004d2b5
 
-# TODO: Standardize on a better approach for how we download models into our images across our packages
 # NOTE: This is checking for a pre-downloaded model file in the local build dir before downloading the model from HuggingFace
 # TODO: Add checksum validation to verify the model in the local build-dir is the model we expect
 COPY scripts/model_download.py scripts/model_download.py
@@ -25,7 +24,6 @@ ENV PATH="/leapfrogai/.venv/bin:$PATH"
 
 # copy the llama-cpp-python build dependencies over
 # NOTE: We are copying to this filename because installing 'optional extras' from a wheel requires the absolute path to the wheel file (instead of a wildcard whl)
-# TODO: This install seems slow. Evaluate options such as copying ALL of the whl's to this image OR vendoring/bundling the dependencies inside of our built .whl
 COPY build/*.whl build/
 COPY build/leapfrogai_api*.whl leapfrogai_api-100.100.100-py3-none-any.whl
 RUN pip install leapfrogai_api-100.100.100-py3-none-any.whl[llama-cpp-python] --no-index --find-links=build/

packages/llama-cpp-python/zarf.yaml

@@ -19,7 +19,7 @@ components:
         namespace: leapfrogai
         localPath: chart
         releaseName: llama-cpp-python-model
-        version: 0.0.1 #TODO: Validate this package version
+        version: 0.0.1
         valuesFiles:
           - "llama-cpp-python-values.yaml"
     images:

packages/text-embeddings/zarf.yaml

@@ -25,7 +25,7 @@ components:
         namespace: leapfrogai
         localPath: chart
         releaseName: text-embeddings-model
-        version: 0.0.1 #TODO: Validate this package version
+        version: 0.0.1
         valuesFiles:
           - "embedding-values.yaml"
     images:

packages/vllm/Dockerfile

@@ -13,16 +13,13 @@ RUN groupadd -g 65532 vglusers && \
 WORKDIR /home/leapfrogai
 
 # grab necesary python dependencies
-# TODO @JPERRY: Get context as to why we are doing this for this Dockerfile but not our other ones
 RUN apt-get -y update \
     && apt-get -y install software-properties-common \
     && add-apt-repository universe \
     && add-apt-repository ppa:deadsnakes/ppa \
     && apt-get -y update
 
 
-# TODO @JPERRY: Install python 3.11.6 specifically by installing it via pyenv...
-# TODO @JPERRY: Determine if we really need to be installing 'git' here
 RUN apt-get -y install python3.11-full
 
 # get deps for vllm compilation and model download
@@ -53,4 +50,4 @@ COPY config.yaml .
 
 EXPOSE 50051:50051
 
-ENTRYPOINT ["python3.11", "-m", "leapfrogai_api.types.cli", "--app-dir=.", "main:Model"]
+ENTRYPOINT ["python3.11", "-m", "leapfrogai_api.types.cli", "--app-dir=.", "main:Model"]

packages/vllm/zarf.yaml

@@ -18,7 +18,7 @@ components:
         namespace: leapfrogai
         localPath: chart
         releaseName: vllm-model
-        version: 0.0.1 #TODO: Validate this package version
+        version: 0.0.1
         valuesFiles:
           - "vllm-values.yaml"
     images:

packages/whisper/Dockerfile

@@ -7,8 +7,6 @@ RUN python3.11 -m venv .venv
 ENV PATH="/leapfrogai/.venv/bin:$PATH"
 
 # download and covnert OpenAI's whisper base
-# TODO @JPERRY: Move this model download to a separate build stage (within its own FROM section)
-#               We don't dirty our .venv with libraries only needed for model download...
 ARG MODEL_NAME=openai/whisper-base
 RUN pip install transformers[torch]
 RUN pip install ctranslate2
@@ -45,4 +43,4 @@ COPY main.py .
 
 EXPOSE 50051:50051
 
-ENTRYPOINT ["python", "-u", "main.py"]
+ENTRYPOINT ["python", "-u", "main.py"]

packages/whisper/main.py

@@ -15,7 +15,6 @@
 
 
 def make_transcribe_request(filename, task, language, temperature, prompt):
-    # TODO @JPERRY does this make sense to be in the function? This seems like something that should be done during initialization
     device = "cuda" if GPU_ENABLED else "cpu"
     model = WhisperModel(model_path, device=device, compute_type="float32")
 

packages/whisper/zarf.yaml

@@ -25,7 +25,7 @@ components:
         namespace: leapfrogai
         localPath: chart
         releaseName: whisper-model
-        version: 0.0.1 #TODO: Validate this package version
+        version: 0.0.1
         valuesFiles:
           - "whisper-values.yaml"
     images:

src/leapfrogai_api/backends/openai/types.py

@@ -86,17 +86,14 @@ class ChatStreamChoice(BaseModel):
     finish_reason: str | None = ""
 
 
-# TODO @JPERRY do we want two distinct response types for stream vs not-stream or do we want the choices to be unioned?
 class ChatCompletionResponse(BaseModel):
     """https://platform.openai.com/docs/api-reference/chat/object"""
 
     id: str = ""
     object: str = "chat.completion"
     created: int = 0
     model: str = ""
-    choices: (
-        list[ChatChoice] | list[ChatStreamChoice]
-    )  # TODO: @JPERRY look into this more, difference between streaming and not streaming
+    choices: list[ChatChoice] | list[ChatStreamChoice]
     usage: Usage | None = None