$ show gabriel-vasile/mimetype
This is a Go library for detecting the MIME type of files using magic number signatures. It implements a hierarchical detection tree rooted at 'application/octet-stream', allowing precise identification of over 200 file formats without any external dependencies. The library provides three detection entry points (byte slice, reader, file path) and supports concurrency via atomic and mutex locks. It includes custom parsers for CSV, JSON, OLE2/CDF, and MP3 frames, all optimized for performance and zero-copy where possible. The codebase is extensively tested with unit tests, fuzzing, and benchmarks, and is actively maintained with CI/CD, code scanning, and dependency automation. It is a standalone library with no runtime dependencies beyond the Go standard library.
The LLM identified 6 files worth highlighting for reuse:
Provides a Bytes type with helper methods for scanning byte slices: line reading, whitespace trimming, pop, peek, pattern matching with flags (CompactWS, IgnoreCase, FullWord). It is a foundational utility used across the library for efficient byte manipulation without allocations. Can be extracted and reused in any Go project needing ad-hoc byte scanning.
A custom, allocation-light JSON parser that validates JSON and optionally matches specific key-value paths to detect GeoJSON, HAR, glTF, and CycloneDX formats. It uses a pool of parser states to reduce allocations and limits recursion depth. Demonstrates non-trivial state-machine based parsing with zero allocation on the fast path.
A memory-efficient CSV reader that counts fields without copying field values. It handles quoted fields, multi-line values, comments, and both LF and CRLF line endings. Tested and fuzzed against the Go standard library csv.Reader. Suitable for extraction into any CSV processing tool that only needs field counts or line structure.
Implements MP3 frame header extraction and validation. It searches for sync bytes (0xFF), validates the header structure, computes frame size from bitrate/sample-rate tables, and confirms frames by checking subsequent headers. Handles padding, ID3v2 tags, and truncated inputs. The algorithm is self-contained and could be reused in audio metadata tools.
Parses OLE2 (CDF) compound document files, detecting embedded documents such as Word, Excel, PowerPoint, MSI, and Outlook messages. It handles fragmented sector chains, short streams, and truncated inputs gracefully. The implementation is permissive and computes SAT entries on the fly to avoid materializing the full SAT. This is a non-trivial binary parser that could be reused for any OLE2-based file inspection.
Provides zip archive inspection without requiring the archive/zip package. It implements a low-level iterator over zip local file headers, extracting entry names efficiently. Also detects Zipflinger virtual entries (used in modern APKs) by checking zero-size entries with no compression. Used for detecting Office Open XML, JAR, APK, and KMZ formats. Extracted as a standalone zip entry scanner.