thisdatedoesnotexist-profile-suggester

A simple API that calculates and returns a list of similar user profiles based on their interests, political views, and religion. The server utilizes TF-IDF (Term Frequency-Inverse Document Frequency) vectorization and cosine similarity to determine profile similarities.

Setup

To set up the profile recommendation system (profile suggester, in this case), follow these steps:

1. Install Dependencies:

   pip install -r requirements.txt

2. Run the Application:

   python app.py

   The server will start at http://localhost:5002.

Description

1. Vectorization and Similarity Calculation

• The script uses the TfidfVectorizer from scikit-learn to convert profile descriptions into numerical vectors.
• Similarity between profiles is calculated using the cosine similarity metric.

2. User Profile Processing

• The preprocess_description function processes user and profile descriptions, considering hobbies, political views, religion, and relationship goals.
• User profiles are enhanced with weights based on political views, favoring similar political perspectives.

3. Finding Similar Profiles

• The find_similar_profiles function takes user data and a list of profiles, calculates similarity scores, and returns suggested profiles.
• Profiles are sorted based on similarity scores, considering political views as a factor.

4. API Endpoint

• The /find-similar-profiles endpoint accepts POST requests with JSON data containing user information and a list of profiles.
• It returns a JSON response with suggested profiles sorted by similarity.

API Usage

• Endpoint: /find-similar-profiles
• Method: POST
• Request Payload:

  {
    "user": {
      "hobbies": [{"name": "hobby1"}, {"name": "hobby2"}],
      "political_view": "center",
      "religion": "agnostic",
      "relationship_goal": {"name": "casual"}
    },
    "profiles": [
      {"id": 1, "hobbies": [...], "political_view": "left", "religion": "atheist", "relationship_goal": {"name": "serious"}},
      ...
    ]
  }

• Response Payload:

  {
    "suggested_profiles": [
      {"id": 2, "profile": {...}, "score": 0.85},
      ...
    ]
  }

Note

• Ensure that the script is executed in a secure environment, especially in production.
• This script is a basic recommendation system and may require further enhancements for a production-ready application.
• Customize the weights, metrics, and additional features as per the application's requirements.
• Handle exceptions and errors gracefully in a production environment.

Dependencies

• Flask
• scikit-learn
• waitress (for production-ready server)