garmin_connector_example/garmin_reqs.plain at main · Codeplain-ai/garmin_connector_example · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
---
description: "Non-functional requirements and environment configuration for the Garmin LLM Chat CLI"
---

***definitions***

- :GarminChatApp: is the core application logic that coordinates data fetching and chat orchestration.

- :GarminClient: is a service that interfaces with the Garmin Connect API.

- :DataFile: is a file that stores the activity data. The default name is "garmin_data.json".

- :LLMClient: is a service that interfaces with Google Gemini.

- :ActivitySummary: is a data structure representing a Garmin activity. :ActivitySummary: has the following attributes:
  - activityId - unique identifier
  - activityName - name of the activity
  - activityType - type of the activity
  - startTimeLocal - date and time of the activity
  - distance - total distance in meters
  - duration - total time in seconds
  - averageHR - average heart rate
  - maxHR - maximum heart rate
  - averageSpeed - average speed

- :LapData: is a data structure representing a single lap within an activity. :LapData: has the following attributes:
  - lapNumber - the sequence number
  - startTime - start time of lap
  - distance - lap distance
  - duration - lap duration
  - averageSpeed - lap speed
  - averageHR - lap heart rate

***implementation reqs***

- :Implementation: must be in Python 3.9 or higher.
- The project must include a "requirements.txt" file listing all dependencies.
- The following environment variables must be used for authentication:
  - GARMIN_EMAIL - The email for Garmin Connect
  - GARMIN_PASSWORD - The password for Garmin Connect
  - GOOGLE_API_KEY - The API key for Google Gemini
- Use the standard "unittest" framework for unit testing.
- The application must handle API errors gracefully by displaying a clear message and exiting with a non-zero status code.
- Sensitive credentials must never be hardcoded and must be read directly from environment variables.

- The reference file [GARMIN_DOCS.md](./docs/GARMIN_DOCS.md) should be used as a reference for the implementation of connecting to Garmin.

- The LLM calls should be make through the google-genai Python package. client = genai.Client(api_key='GOOGLE_API_KEY'); response = client.models.generate_content(model=MODEL_ID,contents='Message text')

***test reqs***

- The venv can be flexible about the versions of dependencies used. This should be reflected in the requirements.txt file.

- The calls to the Garmin or LLM APIs should never be mocked during testing!

- The LLM thinking level should be set to minimal during testing.

- LLM responses may take some time to return. This should be reflected in the test timeouts.

- Use printing to console to debug the application and tests.