Feat: Add free weather module using Meteostat (and fix dependencies)

[AyushAnand413]

Description

Fixes: #47

This PR adds a new, independent feature for fetching rich historical weather data using the free Meteostat library. This solves the core problem of relying on a limited, static precipitation.csv file.

This PR is purely additive and does not change the existing training pipeline. It also fixes critical bugs in the project's setup.

Changes Made

• Added: src/features/weather_api.py - A new module with a function get_weather_for_trip() to get weather for a single, specific time and location.
• Added: src/get_historical_weather.py - A bulk-download script to get the entire 2016 weather dataset for NYC in one command, saving it to data/processed/historical_weather.csv.
• Added: meteostat to requirements.txt.
• Updated: config.py to include a new path for 'historical_weather' in the DATA_PATHS dictionary.
• Updated: The bulk-download script to be more robust, handling the case where 'visibility' data is missing from the API response.

Testing

• Ran unit tests (pnpm test)
• Tested manually (describe below):
  • Test case 1: Test live API module
    • Run: python src/features/weather_api.py
    • Expected Result: Script prints Success! Received data: ...
  • Test case 2: Test bulk-download module
    • Run: python src/get_historical_weather.py
    • Expected Result: Script prints Success! Saved 4368 hourly records to data/processed/historical_weather.csv
  • Test case 3: Test for regressions
    • Run: python main.py
    • Expected Result: The original pipeline runs successfully (proves that my new code did not break any existing logic).

Screenshots

N/A. This is a backend data pipeline enhancement.

Checklist

• My code follows the project's coding standards.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas (via docstrings).
• I have made corresponding changes to the documentation.
• My changes generate no new warnings.
• I have added tests that prove my fix is effective or that my feature works.
• I have verified that no API keys or other secrets are committed. (Meteostat is keyless).
• I have updated .env.example with any new environment variables. (No .env file was needed).

---

💡 How to Evolve the Model with This Data

This PR adds the tools to improve the model. To see the new results, the pipeline needs to be restructured as follows:

1. Get the Data:
   First, run the new bulk-download script one time to create the historical_weather.csv file:

   python src/get_historical_weather.py

2. Modify src/feature_pipe.py:

   • Find the logic that loads precipitation.csv.
   • Remove that logic.
   • Add new logic to load and merge historical_weather.csv:

     # Load the new data
     weather_path = config.DATA_PATHS['historical_weather']
     df_weather = pd.read_csv(weather_path, parse_dates=['datetime_hourly'])
     
     # Create merge key (round pickup time to nearest hour)
     df['datetime_hourly'] = df['pickup_datetime'].dt.floor('h')
     
     # Merge
     df = pd.merge(df, df_weather, on='datetime_hourly', how='left')
     
     # Fill any missing values
     for col in ['temp', 'humidity', 'wind_speed']:
         if col in df.columns:
             df[col] = df[col].fillna(df[col].mean())

3. Modify config.py:

   • Find the FEATURE_COLUMNS dictionary.
   • Remove the 'precipitation' line.
   • Add the new weather features:

     'weather': ['temp', 'humidity', 'wind_speed']

4. Re-Run and Compare:
   Finally, run python main.py. The pipeline will re-build the features and re-train the models. The new RMSE scores can then be compared to the old ones to see the improvement.

[netlify]

✅ Deploy Preview for gopredict ready!

Name	Link
🔨 Latest commit	03922a3
🔍 Latest deploy log	https://app.netlify.com/projects/gopredict/deploys/68f7ef636d585b000861662d
😎 Deploy Preview	https://deploy-preview-49--gopredict.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[netlify]

✅ Deploy Preview for storied-fudge-a7595c canceled.

Name	Link
🔨 Latest commit	03922a3
🔍 Latest deploy log	https://app.netlify.com/projects/storied-fudge-a7595c/deploys/68f7ef63157a8600087ab875

[AyushAnand413]

@harshitaphadtareI have made a PR please read the description for a full breakdown of the changes and testing instructions.
Also please add hacktoberfest label before merging !

[AyushAnand413]

@harshitaphadtare Please reply!

[harshitaphadtare]

hey @AyushAnand413 sorry for the delay. loved the way you implemented it! Thankyou for contributing :)