Skip to content

API_DOC.md

Latest commit

 

History

History
58 lines (41 loc) · 1.66 KB

API_DOC.md

File metadata and controls

58 lines (41 loc) · 1.66 KB
Raw

API — v1.0.0

Base URL: http://localhost:8000 · Interactive docs: /docs (Swagger), /redoc

GET /

Welcome message only.

{ "message": "Welcome to the Text Sentiment Analysis API" }

POST /predict/

Predict sentiment label.

Request

{ "text": "This is a sample text for sentiment analysis." }

Response

{
  "cleaned_text": "this is a sample text for sentiment analysis",
  "sentiment": "neutral"
}

Note: server-side this calls preprocess_text() once; the cleaned text is passed through unchanged to classify_sentiment() — one preprocessing pass per request (DECISIONS.md #3, fixed v1.0.1).

POST /predict-prob/

Same as /predict/, plus raw class probabilities.

Response

{
  "cleaned_text": "this is a sample text for sentiment analysis",
  "sentiment_prob": [0.21, 0.58, 0.21]
}

Order: [negative, neutral, positive]. Values reflect the model's single softmax output (DECISIONS.md #1, fixed v1.0.1). Calibration against human-labeled ground truth is still unverified — training labels are TextBlob polarity heuristics (see MODELS.md #3, DECISIONS.md #13).

POST /clean/

Preprocessing only, no model call.

Response

{ "cleaned_text": "this is a sample text for sentiment analysis" }

Schema notes

  • interfaces/text_sentiment_interface.py migrated to json_schema_extra={"example": ...} under pydantic 2.4.2 (DECISIONS.md #8, fixed v1.0.1). Swagger examples confirmed rendering.
  • No error responses documented (no 4xx/5xx schemas) — FastAPI default validation errors only.
  • No versioning in the URL path (/v1/predict etc.) — relevant if 2.0 changes response shape.