fix: inference-mock code review changes

Signed-off-by: Emilio Garcia <i.am.emilio@gmail.com>

Author: iamemilio

.github/workflows/inference-mock.yml

@@ -11,19 +11,25 @@ on:
 
 jobs:
   inference-mock-unit-tests:
-    runs-on: ubuntu:latest
+    runs-on: ubuntu-latest
     strategy:
       matrix: 
         python-version: ["3.11"]
     steps:
+      - uses: actions/checkout@v4
+        with:
+          sparse-checkout: |
+            actions/inference-mock
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
           python-version: ${{ matrix.python-version }}
       - name: Install dependencies
+        working-directory: actions/inference-mock
         run: |
           python -m pip install --upgrade pip
           pip install -r requirements.txt
       - name: Run Unit Tests
+        working-directory: actions/inference-mock/test
         run: |
-          python -m unittest test/test.py
+          python -m unittest test.py

.pylintrc

@@ -627,7 +627,7 @@ missing-member-max-choices=1
 mixin-class-rgx=.*[Mm]ixin
 
 # List of decorators that change the signature of a decorated function.
-signature-mutators=unittest.mock.patch,unittest.mock.patch.object
+signature-mutators=unittest.mock.patch,unittest.mock.patch.object,click.decorators.option
 
 
 [VARIABLES]

README.md

@@ -21,7 +21,7 @@ Below is a list of the in-house GitHub actions stored in this repository:
 | [launch-ec2-runner-with-fallback](./actions/launch-ec2-runner-with-fallback/launch-ec2-runner-with-fallback.md) | Used launch an EC2 instance in AWS, either as a spot instance or a dedicated instance. If your preferred availability zone lacks availability for your instance type, "backup" availability zones will be tried. | <ul><li>Insufficient capacity in AWS (i.e., AWS lacks availablility for your desired EC2 instance type in your preferred availability zone)</li><li>Cost savings (i.e., You want to try launching your EC2 runner as a spot instance first)</li></ul> |
 | [validate-notebooks](./actions/launch-ec2-runner-with-fallback/launch-ec2-runner-with-fallback.md) | Used to validate `.ipynb` files | <ul><li>I maintain a collection of `.ipynb` files and run ci jobs to test them. I  would like to quickly verify that the files are formatted correctly before spinning up more complex or expensive CI jobs to test those notebooks.</li></ul>
 | [update-constraints](./actions/update-constraints/update-constraints.md) | Used to update `constraints-dev.txt` file | <ul><li>I pin all project dependencies in CI using `constraints-dev.txt` file. I would like to monitor new dependency releases and periodically post PRs to move pins forward.</li></ul>
-| [inference-mock](./actions/inference-mock/README.md) | Used to mock LLM calls | <ul><li>I have a notebook that I want to test that makes an LLM call, but I dont need to rely heavily on an LLM and don't need to run a real inference server for this test.</li></ul> |
+| [inference-mock](./actions/inference-mock/README.md) | Used to mock LLM calls | <ul><li>I have a notebook that I want to test that makes an LLM call, but I don't need to rely heavily on an LLM and don't need to run a real inference server for this test.</li></ul> |
 
 ## ❓ How to Use One or More In-House GitHub Actions
 

actions/inference-mock/README.md

@@ -3,7 +3,7 @@
 ## Overview
 
 Inference Mock is a tool that creates a flask server that runs as a background process. OpenAI comptabile calls can be made to its completions API.
-Based on how the server is configured, it will send a set of programmed resposnes back.
+Based on how the server is configured, it will send a set of programmed responses back.
 
 ## When to Use it?
 
@@ -29,7 +29,7 @@ port: 11434
 #   - response: passing only a response is an `Always` match strategy. If no other strategy has matched yet, this will always be a positive match.
 #
 # note: the strategies are executed in the order listed, and the first succesful match is accepted. If you start with an `Always` strategy, its 
-# response will be the only resposne returned.
+# response will be the only response returned.
 matches:
 
     # this is an example of a `contains` strategy. If the prompt contains the substrings, it returns the response.

actions/inference-mock/app.py

@@ -23,12 +23,15 @@ def completions():
     if not data or "prompt" not in data:
         raise exceptions.BadRequest("prompt is empty or None")
 
-    prompt_debug_str = data["prompt"][:90] + "..."
+    prompt = data.get('prompt')
+    prompt_debug_str = prompt
+    if len(prompt) > 90:
+        prompt_debug_str = data["prompt"][:90] + "..."
+
     app.logger.debug(
         f"{request.method} {request.url} {data['model']} {prompt_debug_str}"
     )
 
-    prompt = data.get("prompt")
     chat_response = strategies.find_match(
         prompt
     )  # handle prompt and generate correct response
@@ -52,7 +55,7 @@ def start_server(config):
     # get config
     yaml_data = yaml.safe_load(config)
     if not isinstance(yaml_data, dict):
-        raise ValueError("config file format is invalid")
+        raise ValueError(f"config file {config} must be a set of key-value pairs")
 
     conf = Config(**yaml_data)
 
@@ -72,4 +75,4 @@ def start_server(config):
 
 
 if __name__ == "__main__":
-    start_server()  # pylint: disable=no-value-for-parameter
+    start_server() # pylint: disable=no-value-for-parameter

actions/inference-mock/matching/matching.py

@@ -6,7 +6,7 @@
 
 class Match(Protocol):
     """
-    Match represnts a single prompt matching
+    Match represents a single prompt matching
     strategy. When a match is successful,
     the response is what should be returned.
     """
@@ -69,8 +69,8 @@ def to_match(pattern: dict) -> Match:
         )
     if "contains" in pattern:
         return Contains(**pattern)
-    else:
-        return Always(**pattern)
+    
+    return Always(**pattern)
 
 
 class Matcher:

actions/inference-mock/test/__init__.py

@@ -4,7 +4,6 @@
 
 
 class TestAlways(unittest.TestCase):
-    
     # match on any prompt
     def test_always(self):
         expect_response = "expected response"
@@ -23,23 +22,22 @@ def test_always_empty_prompt(self):
 
 
 class TestContains(unittest.TestCase):
-
     def test_contains(self):
         expect_response = "expected response"
         prompt = "example prompt"
         match_on = ["example"]
         contains = Contains(match_on, expect_response)
         actual_response = contains.match(prompt)
         self.assertEqual(actual_response, expect_response)
-    
+
     def test_contains_many(self):
         expect_response = "expected response"
         prompt = "a much longer example prompt so we can match on many substring elements of this string"
         match_on = ["example", "many substring elements", "match on"]
         contains = Contains(match_on, expect_response)
         actual_response = contains.match(prompt)
         self.assertEqual(actual_response, expect_response)
-    
+
     # if any substrings don't match, return None
     def test_contains_mismatch(self):
         response = "expected response"
@@ -48,7 +46,7 @@ def test_contains_mismatch(self):
         contains = Contains(match_on, response)
         actual_response = contains.match(prompt)
         self.assertIsNone(actual_response)
-    
+
     # reject empty prompts
     def test_contains_empty(self):
         response = "expected response"
@@ -60,7 +58,6 @@ def test_contains_empty(self):
 
 
 class TestMatcher(unittest.TestCase):
-
     def test_to_contains(self):
         response = 'I am a response'
         substr = ["a", "b", "c"]
@@ -75,7 +72,7 @@ def test_to_always(self):
         always = to_match(always_pattern)
         self.assertIsInstance(always, Always)
         self.assertEqual(always.response, response)
-    
+
     def test_to_invalid(self):
         response = 'I am a response'
         invalid_pattern = {'banana':'foo', 'response': response}