NLP Analysis

Endpoint: POST /nlp-analysis

Request Parameters

File Upload

• file (required): The file containing the text data to be analyzed. Supported formats include CSV.
• text_column (required): The name of the column containing the text data.
• analysis_type (optional): The specific type of analysis to perform (ner, similarity, language, classification, sentiment, keywords, topics, summary). If omitted, all analyses will be performed.
• num_keywords (optional, default=5): The number of keywords to extract during keyword extraction.
• num_topics (optional, default=3): The number of topics for topic modeling.
• summary_length (optional, default=3): The number of sentences in the text summary.

Example Request

curl -X POST "http://localhost:8000/nlp-analysis" \
-F "file=@data.csv" \
-F "text_column=review" \
-F "analysis_type=sentiment"

Analysis Components

1. Named Entity Recognition (NER): Identifies entities like organizations, people, locations, dates, and products.
2. Text Similarity: Calculates cosine similarity between text samples.
3. Language Detection: Detects the language of the text data.
4. Text Classification: Classifies text as positive, negative, or neutral based on sentiment scores.
5. Sentiment Analysis: Uses VADER to analyze sentiment scores (positive, negative, neutral, compound).
6. Keyword Extraction: Extracts keywords based on term frequency.
7. Topic Modeling (LDA): Discovers topics and related keywords from the text.
8. Text Summarization: Generates a brief extractive summary of the text data.

Example Response

{
  "results": {
    "ner": {
      "ORG": ["Google"],
      "PERSON": ["John"],
      "GPE": ["France"],
      "DATE": ["2023"],
      "PRODUCT": ["iPhone"]
    },
    "similarity": [[1.0, 0.5], [0.5, 1.0]],
    "language": ["en", "fr"],
    "classification": ["positive", "negative"],
    "sentiment": [
      {"neg": 0.1, "neu": 0.7, "pos": 0.2, "compound": 0.25},
      {"neg": 0.4, "neu": 0.5, "pos": 0.1, "compound": -0.4}
    ],
    "keywords": ["product", "quality", "service"],
    "topics": [
      {"topic": 1, "keywords": ["product", "service", "quality"]},
      {"topic": 2, "keywords": ["price", "value", "cost"]}
    ],
    "summary": "The product quality is excellent, but the price is high."
  }
}