102 projects
cjm-transcription-tui
Transcription-workflow TUI driver: source/folder selection, candidate (capability, model)-instance setup from manifest config schemas, and per-segment side-by-side transcriber comparison to pick a run's lightweight/accuracy pair — then hands off to cjm-transcription-core's headless pipeline. Born on-graph.
cjm-transcription-core
A frontend-agnostic core for the audio transcription workflow — composes isolated capability workers (audio conversion, VAD segmentation, batch transcription, persistence) into a headless pipeline, with a CLI as its first driver.
cjm-transcript-graph-schema
Audio-transcript layer schema for context graphs: Source, AudioSegment, Transcript, and Segment nodes with deterministic identity tuples and graph-node mapping, shared by the transcription, decomposition, and correction workflow cores.
cjm-transcript-correction-tui
Transcript-correction TUI driver: keyboard-first terminal review surface (cursor-windowed segment walk, VAD-chunk audio auto-play, fidelity + boundary-shift corrections, agents-propose/humans-confirm triage) over cjm-transcript-correction-core's operation vocabulary. The first born-on-graph library of the self-hosting graph arc.
cjm-substrate
A dependency-isolated capability-composition runtime: heterogeneous tools run in their own environments behind a uniform HTTP/JSON boundary; the host composes them into workflows with resource-aware scheduling and threads provenance through a context graph.
cjm-transcript-decomp-tui
Decomp-batch TUI driver: browse transcription-core run manifests, queue an ordered multi-run batch with per-run authoritative-text picks, then hand off to cjm-transcript-decomp-core's headless batch runner (one capability stack for the whole batch). Born on-graph.
cjm-substrate-tui-kit
Shared TUI components for cjm-substrate workflow frontends: manifest-driven config forms from capability config schemas, cursor-windowed viewport math, sidecar state store, audio chunk playback, and repaint governance — extracted demand-driven from the transcription and transcript-correction TUIs. Born on-graph.
cjm-transcript-decomp-core
A frontend-agnostic core for the transcript decomposition workflow — composes isolated capability workers (forced alignment, VAD, graph storage) into a headless pipeline that decomposes transcription run manifests into a VAD-aligned context-graph spine with traceable provenance, and a CLI as its first driver.
cjm-transcript-correction-core
A frontend-agnostic core for the transcript correction workflow — the first downstream graph-extension core; composes the graph-storage capability worker into a headless pipeline that applies unified, non-destructive corrections (text, punctuation, segmentation) as a supersede-able overlay on a committed decomposition spine, recomputes the review worklist from deterministic signals, and exposes a CLI as its first driver.
cjm-capability-whisper
An OpenAI Whisper speech-to-text capability for the cjm-substrate runtime that provides local transcription with configurable model selection and parameter control.
cjm-context-graph-projection
Projection and navigation core for context graphs: bounded, ranked, provenance-carrying reads (schema / state / relevance / show) over any cjm-substrate context graph, with a CLI driver. The agent-facing read layer of the self-hosting graph arc.
cjm-context-graph-layer
Domain-neutral graph-aware layer for context graphs: deterministic node identity, spine and overlay grammar, spine-edit operations with effective-view projection, supersession resolution, idempotent emission/extension, and provenance-by-declaration.
cjm-context-graph-primitives
The dep-light data-primitives library for the cjm context-graph ecosystem — defines the shared data nouns that workflow cores, graph-storage adapters, provenance bundles, and the graph-aware composition layer all speak: structured resource locators with canonical URI rendering, content-hash-primary SourceRef provenance references, atomic typed content slices, GraphNode/GraphEdge/GraphContext containers, and the structured typed query expression executed by graph-storage adapters.
cjm-substrate-hf-utils
Hugging Face Hub helpers for cjm-substrate capabilities: cache-config mixin, progress-reporting snapshot downloads, and typed-OOM model loading.
cjm-substrate-torch-utils
PyTorch helpers for cjm-substrate capabilities: GPU memory release, typed CUDA-OOM handling, and device selection.
cjm-notebook-decompose-core
A Jupyter/nbdev notebook decomposition core for context graphs: a notebook IS a CodeModule whose authored source is an ordered sequence of VERBATIM cells (the lossless source substrate), composing the python decomposer (code cells -> symbols) and the markdown decomposer (markdown cells -> prose) with interleaving edges. The first step toward graph-as-source-of-truth (notebook -> graph).
cjm-markdown-decompose-core
A Markdown decomposition core for context graphs: parses Markdown content (frontmatter, structure, and [[wiki-links]]) into provenance-carrying graph nodes and edges. First source = the ecosystem's own memory/decision files; generalizes to any Markdown corpus.
cjm-python-decompose-core
A Python source-code decomposition core for context graphs: parses .py modules (docstrings, imports, and the symbol tree of functions/classes/methods with their calls) into provenance-carrying graph nodes and edges that co-reside with the markdown source-type. First source = the ecosystem's own code; generalizes to any Python corpus.
cjm-dev-graph-schema
Development/decision-provenance schema for context graphs: Decision, Fact-slot/Assertion, Evidence, Thread, Session, Procedure, and Entity node kinds with deterministic identity and overlay/reasoning edges — the dev-domain sibling to cjm-transcript-graph-schema for graphing a project's own evolution.
cjm-vad-adapter-interface
Typed voice-activity-detection task-adapter interface — VADAdapter ABC + GenericVADAdapter (cache/persist bookends around a pure-compute tool), the VADToolProtocol, and VAD persistence helpers. The VADResult data noun lives in cjm-capability-primitives.
cjm-source-separation-adapter-interface
Typed source-separation (audio-preprocessing) task-adapter interface — SourceSeparationAdapter ABC + GenericSourceSeparationAdapter (cache/persist bookends around a pure-compute tool that produces a vocals-isolated audio artifact), the SourceSeparationToolProtocol, and source-separation persistence helpers.
cjm-media-processing-adapter-interface
Typed media-processing task-adapter interface — the multi-method MediaProcessingAdapter ABC + GenericMediaProcessingAdapter (cache/persist bookends around a pure-compute tool: convert, segment_audio, extract_audio, plus an uncached get_info probe) + the MediaProcessingToolProtocol. Result DTOs live in cjm-capability-primitives.
cjm-transcription-adapter-interface
Typed transcription task-adapter interface — TranscriptionAdapter ABC + GenericTranscriptionAdapter (cache/persist bookends around a pure-compute tool), the TranscriptionToolProtocol, and transcription persistence helpers. The TranscriptionResult data noun lives in cjm-capability-primitives (re-exported here as a compat shim).
cjm-forced-alignment-adapter-interface
Typed forced-alignment task-adapter interface — ForcedAlignmentAdapter ABC, ForcedAlignResult/ForcedAlignItem wire DTOs, and forced-alignment persistence helpers.
cjm-provenance-bundle
Generic provenance bundle envelope: typed hash-linked stage chains, payload staging and relocation, reference-vs-copy dispositions, integrity/chain validation, and atomic packaging for sharing workflow outputs with verifiable provenance.
cjm-graph-storage-adapter-interface
Typed graph-storage task adapter interface: the repository-style contract (CRUD, typed query execution, introspection) between graph-storage tool capabilities and the substrate.
cjm-capability-primitives
The dep-light primitives library for the cjm capability ecosystem — defines the shared cross-task result DTOs (data nouns) that tool capabilities emit and task adapters, workflow cores, and the composition layer all consume, so a pure-compute tool capability depends only on its task's data noun, never on the adapter machinery.
cjm-transcript-segment-align
FastHTML dual-column text segmentation & VAD alignment UI for transcript decomposition workflows with forced alignment-based text splitting for aligning text segments with VAD chunks.
cjm-transcription-plugin-system
A flexible plugin system for audio transcription intended to make it easy to add support for multiple backends.
cjm-transcription-plugin-qwen3-forced-aligner
A Qwen3-based forced alignment plugin for the cjm-transcription-plugin-system that provides word-level audio-text alignment using Qwen/Qwen3-ForcedAligner-0.6B with progress reporting.
cjm-system-monitor-nvidia
An NVIDIA GPU monitoring plugin for the cjm-plugin-system that provides real-time hardware telemetry via nvitop.
cjm-media-plugin-demucs
Demucs v4 audio source separation plugin for the cjm-plugin-system that provides vocals extraction for removing background noise and music from speech audio.
cjm-transcription-plugin-whisper
OpenAI Whisper plugin for the cjm-transcription-plugin-system library - provides local speech-to-text transcription with configurable model selection and parameter control.
cjm-transcription-plugin-voxtral-hf
Mistral Voxtral plugin for the cjm-transcription-plugin-system library - provides local speech-to-text transcription through 🤗 Transformers with configurable model selection and parameter control.
cjm-media-plugin-silero-vad
A local, persistent Voice Activity Detection (VAD) worker for the cjm-plugin-system that provides high-accuracy speech segmentation using Silero VAD with SQLite result caching.
cjm-media-plugin-ffmpeg
FFmpeg-based media processing plugin for the cjm-plugin-system that provides audio extraction, segmentation, format conversion, and segment extraction.
cjm-graph-plugin-sqlite
A local, file-backed Context Graph worker for the cjm-plugin-system that implements graph storage, traversal, and querying using SQLite.
cjm-transcription-plugin-gemini
Google Gemini API plugin for the cjm-transcription-plugin-system library - provides speech-to-text transcription with configurable model selection and parameter control.
cjm-transcription-plugin-voxtral-vllm
Mistral Voxtral plugin for the cjm-transcription-plugin-system library - provides local speech-to-text transcription through vLLM with configurable model selection and parameter control.
cjm-transcript-review
FastHTML Reader View component for verifying transcript segmentation with audio playback and context graph commit.
cjm-transcript-verify
A FastHTML verification step for transcript decomposition workflows that queries the context graph database to display integrity checks and sample inspection.
cjm-transcript-workflow-management
A FastHTML management interface for context graph documents to list, inspect, delete, and import/export graph spines produced by transcript decomposition workflows.
cjm-text-plugin-nltk
A local, NLTK-based text processing worker for the cjm-plugin-system that provides sentence splitting and tokenization with character-level span tracking.
cjm-transcription-audio-segment
FastHTML audio segmentation component for transcription workflows — VAD analysis, configurable boundary-midpoint segmentation, audio splitting at speech boundaries, and spot-check verification.
cjm-transcription-source-select
FastHTML source selection component for transcription workflows that provides local file browsing with audio/video preview, selection ordering, and integrated audio extraction from video sources.
cjm-transcript-vad-align
FastHTML VAD alignment component for transcript decomposition with Silero VAD audio chunking, card stack navigation with audio playback, and Web Audio API integration.
cjm-graph-plugin-system
Defines the standardized interface and data structures for Context Graph plugins, enabling the semantic linking, decomposition, and enrichment of multi-modal content.
cjm-transcript-segmentation
FastHTML text segmentation component for transcript decomposition with NLTK sentence splitting, interactive split/merge UI with token selector, and card stack navigation.
cjm-hf-plugin-utils
Hugging Face Hub helpers for cjm-plugin-system plugins: cache-config mixin, progress-reporting snapshot downloads, and typed-OOM model loading.
cjm-infra-plugin-system
Standardized interface and data structures for infrastructure monitoring plugins in the cjm-plugin-system ecosystem.
cjm-text-plugin-system
Defines standardized interfaces and data structures for text processing plugins, enabling modular NLP operations like sentence splitting, tokenization, and chunking within the cjm-plugin-system ecosystem.
cjm-fasthtml-job-monitor
A FastHTML component library for monitoring plugin job execution with progress tracking, log tailing, resource usage, and cancellation via a tabbed modal with content overlay.
cjm-media-plugin-system
Defines standardized interfaces and data structures for media analysis (VAD, Scene Detection) and processing (FFmpeg, Conversion) plugins within the cjm-plugin-system ecosystem.
cjm-transcript-source-select
FastHTML source selection component for transcript decomposition workflows, with federated database browsing, drag-drop ordering, and keyboard navigation.
cjm-torch-plugin-utils
PyTorch helpers for cjm-plugin-system plugins: GPU memory release, typed CUDA-OOM handling, and device selection.
cjm-plugin-system
Process-isolated plugin system with resource-aware scheduling, configuration validation, and lifecycle management for extensible Python applications.
cjm-graph-domains
A shared library of Pydantic models and standard schemas for the cjm-graph-plugin-system ecosystem, enabling type-safe domain definitions across plugins and UIs.
cjm-nbdev-utils
Miscellaneous utilities for nbdev projects.
cjm-nbdev-overview
Automated documentation and visualization tools for nbdev projects.
cjm-fasthtml-workflow-session-management
A workflow-agnostic FastHTML library for managing multiple in-progress workflow sessions with list, resume, create, delete, and rename sessions stored in a cjm-workflow-state state database.
cjm-fasthtml-file-browser
Interactive file system navigation and selection UI components for FastHTML applications with extensible provider support.
cjm-fasthtml-web-audio
A reusable Web Audio API manager for FastHTML applications with multi-buffer support, parallel decode, and card-stack-compatible playback.
cjm-fasthtml-virtual-collection
FastHTML virtualized collection rendering with discrete navigation, custom scrollbar, table layout, and cell-level HTMX updates.
cjm-fasthtml-card-stack
A fixed-viewport card stack component for FastHTML with keyboard navigation, scroll-to-nav, configurable focus position, and HTMX-driven OOB updates.
cjm-fasthtml-token-selector
A token-level interactive text selector for FastHTML with gap, word, and span selection modes, custom key repeat navigation, and HTMX state sync.
cjm-fasthtml-sortable-queue
A reusable FastHTML component for Sortable.js-enhanced ordered queue panels with drag-and-drop reorder, keyboard navigation, and HTMX-powered mutations.
cjm-fasthtml-step-progress
A compact, theme-aware step progress bar component for FastHTML multi-step workflows.
cjm-fasthtml-viewport-fit
Reusable viewport height measurement utility for FastHTML applications that dynamically fits element height to remaining viewport space with resize and HTMX integration.
cjm-fasthtml-media-gallery
FastHTML UI components for displaying media file collections with type-specific previews, playback, and pagination.
cjm-fasthtml-interactions
Reusable user interaction patterns for FastHTML applications, providing the StepFlow multi-step wizard and AsyncLoadingContainer for asynchronous content loading.
cjm-fasthtml-virtual-scrollbar
FastHTML virtual scrollbar component with drag-to-scrub, click-to-jump, and position indication for discrete-navigation containers.
cjm-fasthtml-keyboard-navigation
A declarative keyboard navigation framework for FastHTML applications with multi-zone focus management, mode switching, and HTMX integration.
cjm-fasthtml-app-core
Core utilities and reusable patterns for FastHTML applications including page layouts, HTMX request handling, navbar, and confirm modal components.
cjm-fasthtml-design-system
Named UI role recipes (panels, chrome, insets) for FastHTML applications — a stable abstraction layer over DaisyUI and Tailwind primitives.
cjm-fasthtml-jsonschema
Library for generating FastHTML user interfaces from JSON Schema configurations.
cjm-fasthtml-daisyui
Python-native daisyUI v5 component and utility class builder providing dynamic class generation with full abstraction for FastHTML projects.
cjm-fasthtml-tailwind
Python-native Tailwind CSS v4 utility builder providing dynamic class generation with full abstraction for FastHTML projects.
cjm-workflow-state
SQLite-backed workflow state persistence and undo history for multi-step human-in-the-loop workflows.
cjm-demucs-v4
Inference-only fork of Demucs v4 that provides audio source separation with TorchCodec replacing torchaudio I/O.
cjm-ffmpeg-utils
Audio/video processing utilities with FFmpeg: extraction, conversion, metadata handling, and progress bars.
cjm-source-provider
Protocol and data types for querying content sources across decomposition workflows.
cjm-file-discovery
Pure Python library for batch file scanning with configurable filtering, metadata extraction, caching, and extensible provider support.
cjm-fasthtml-lucide-icons
Lucide icons as FastHTML SVG components with category/tag registry and factory API.
cjm-fasthtml-workflow-transcription-single-file
A self-contained single-file transcription workflow for FastHTML applications.
cjm-fasthtml-settings
A drop-in schema-based configuration system for FastHTML applications with automatic UI generation, sidebar navigation, and persistent storage.
cjm-fasthtml-sse
Real-time Server-Sent Events (SSE) and HTMX integration library for FastHTML with cross-tab synchronization support.
cjm-error-handling
Structured error handling with context propagation, user-friendly messaging, and serialization support for reusable Python libraries.
cjm-transcription-utils
Miscellaneous utilities for helping with audio transcription.
cjm-tqdm-capture
Intercept and capture progress information from tqdm progress bars via callbacks without modifying existing code.
cjm-nbdev-docments
Tool for validating and updating documentation in nbdev projects for use with the fastcore.docments module.
cjm-fasthtml-sysmon
FastHTML components and utilities for real-time system resource monitoring with SSE-based live updates.
cjm-fasthtml-byok
Secure API key management for FastHTML applications with encrypted storage, session/database persistence, and built-in UI components.
cjm-torchvision-tfms
Some custom Torchvision tranforms.
cjm-yolox-pytorch
A PyTorch implementation of the YOLOX object detection model based on the mmdetection implementation.
cjm-pytorch-utils
Some utility functions for working with PyTorch.
cjm-pil-utils
Some PIL utility functions I frequently use.
cjm-byte-track
A standalone Python implementation of the ByteTrack multi-object tracker based on the official implementation.
cjm-psl-utils
Some utility functions using the Python Standard Library.
cjm-pandas-utils
Some utility functions for working with Pandas.
cjm-diffusers-utils
Some utility functions I frequently use with 🤗 diffusers.
cjm-kaggle-utils
Some utility functions for working with the Kaggle API.
cjm-parallel-utils
Utility functions for parallel operations.
