Contains isolated utility functions for processing HAR (HTTP Archive) files. Includes functions to parse HAR entries into Request objects, build URL-to-request mappings, extract URL details while filtering noise (tracking domains, media files), and parse cookie files. Demonstrates non-obvious skill in properly handling different content types, filtering out common tracking/infrastructure noise to reduce LLM processing load, and robustly parsing various request/response formats. Could be extracted and used in any project needing to analyze browser network traffic.

Provides sophisticated code generation capabilities from a dependency graph of HTTP requests. Includes functions to visualize DAGs, generate Python functions for each request based on its attributes (handling different response types like JSON, binary files, HTML with appropriate parsing strategies), and aggregate generated functions into a runnable sequence. Shows non-obvious skill in adaptive code generation - dynamically choosing between JSON path extraction, regex parsing for HTML, or file saving for binary responses based on content type. The generated code handles dependencies properly by passing extracted values between functions. Could be extracted as a general DAG-to-executable-code converter.

Defines a self-contained Request class that represents HTTP requests with methods to convert to and from curl command strings. Handles various HTTP methods, headers, query parameters, and request bodies (including JSON and form data). Properly manages content-type headers and body serialization. Demonstrates clean encapsulation of HTTP request details with thoughtful handling of edge cases. Could be extracted and used in any project needing to work with HTTP requests in curl format or needing to serialize/deserialize HTTP requests.