| name | text-summarizer |
| description | Advanced text summarization and translation service with async processing capabilities. Use when users need to: (1) Summarize large text datasets from CSV/DataFrame, (2) Batch process text documents with length-based intelligent routing, (3) Translate shorter text content, (4) Configure custom processing parameters for text analysis workflows, (5) Handle text processing tasks requiring async operations for performance. |
Text Summarizer
Intelligent text processing service that automatically determines whether to summarize or translate content based on configurable length thresholds, with async batch processing capabilities.
Core Functionality
Use scripts/text_summarizer.py for bulk text processing operations:
- Smart routing: Texts above length threshold ? summarization; texts below ? translation
- Async processing: Efficient batch operations with concurrent API calls
- Flexible input: Process pandas DataFrames or CSV files
- Configurable: Adjustable length thresholds, model selection, temperature settings
Quick Start
Process a DataFrame with default settings:
uv run scripts/text_summarizer.py --input data.csv --text-column content --output results.csv
Advanced Configuration
Custom length threshold (default: 500 words):
uv run scripts/text_summarizer.py --input data.csv --length 1000 --text-column review
Model selection (default: gpt-4.1):
uv run scripts/text_summarizer.py --input data.csv --model gpt-4o-mini --temperature 0.3
API Reference
For programmatic usage and advanced customization, see API_REFERENCE.md for complete NotesSummarizer class documentation.
Error Handling
- Missing dependencies: Ensure LangChain, pandas, and OpenAI libraries are installed
- API limits: Script handles retries automatically with exponential backoff
- File access: Verify input file paths and write permissions for output files
- Column names: Specify correct text column name with
--text-column parameter
Environment Requirements
- Python 3.11+
- OpenAI API key configured
- Required packages:
langchain, pandas, tqdm, numpy
