m3tam3re/nixpkgs
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: add auto-language detection support to stt-ptt

Browse Source 
- Add STT_LANGUAGE environment variable (default: auto)
- Modify whisper-cli call to use -l flag with language code
- Add language option to HM module (enum with 23 common languages + auto)
- Update help text to document STT_LANGUAGE variable
This commit is contained in:
m3tm3re
2026-01-10 09:36:21 +01:00
parent 3933e380c6
commit 9f284a8c21
3 changed files with 55 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
.beads/issues.jsonl
View File

@@ -2,6 +2,6 @@
{"id":"nixpkgs-5ml","title":"Update rofi-project-opener script to pass args to opencode","description":"Modify pkgs/rofi-project-opener script to read the new config format and launch opencode with: 'opencode \u003cdirectory\u003e \u003carguments\u003e' instead of just 'opencode' in the directory.","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-05T12:45:49.748958951+01:00","created_by":"m3tam3re","updated_at":"2026-01-05T12:49:40.800083615+01:00","closed_at":"2026-01-05T12:49:40.800083615+01:00","close_reason":"Script updated to read JSON config and pass args to opencode","dependencies":[{"issue_id":"nixpkgs-5ml","depends_on_id":"nixpkgs-w3u","type":"parent-child","created_at":"2026-01-05T12:46:02.338350208+01:00","created_by":"m3tam3re"},{"issue_id":"nixpkgs-5ml","depends_on_id":"nixpkgs-sys","type":"blocks","created_at":"2026-01-05T12:46:04.966269033+01:00","created_by":"m3tam3re"}]}
{"id":"nixpkgs-e2u","title":"Change projectDirs from list to attrset with path and args","description":"Change projectDirs option type from 'types.listOf types.str' to an attrset like:\n\nprojectDirs = {\n nixpkgs = { path = \"~/p/NIX/nixpkgs\"; args = \"--agent Planner-Sisyphus\"; };\n myproject = { path = \"~/dev/myproject\"; }; # args optional\n};\n\nMust maintain backward compatibility consideration.","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-05T12:45:48.6992807+01:00","created_by":"m3tam3re","updated_at":"2026-01-05T12:48:12.178120719+01:00","closed_at":"2026-01-05T12:48:12.178120719+01:00","close_reason":"Changed projectDirs from listOf str to attrsOf submodule with path+args","dependencies":[{"issue_id":"nixpkgs-e2u","depends_on_id":"nixpkgs-w3u","type":"parent-child","created_at":"2026-01-05T12:46:00.515400521+01:00","created_by":"m3tam3re"}]}
{"id":"nixpkgs-hrh","title":"projectDirs on rofi-project-switcher should be exported to users home-manager environment","status":"closed","priority":2,"issue_type":"feature","created_at":"2026-01-05T11:46:43.640224459+01:00","created_by":"m3tam3re","updated_at":"2026-01-05T12:49:43.584087541+01:00","closed_at":"2026-01-05T12:49:43.584087541+01:00","close_reason":"Feature complete: config exported as ROFI_PROJECT_OPENER_CONFIG env var","dependencies":[{"issue_id":"nixpkgs-hrh","depends_on_id":"nixpkgs-w3u","type":"blocks","created_at":"2026-01-05T12:46:05.867959608+01:00","created_by":"m3tam3re"}]}
{"id":"nixpkgs-p79","title":"feat: Add auto-language detection support to stt-ptt package","description":"## Current State\n- **Package**: `/home/m3tam3re/p/NIX/nixpkgs/pkgs/stt-ptt/default.nix`\n- **Module**: `/home/m3tam3re/p/NIX/nixpkgs/modules/home-manager/cli/stt-ptt.nix`\n- **Current whisper-cli call** (line 55): `\"$WHISPER\" -m \"$STT_MODEL\" -f \"$AUDIO\" -np -nt 2\u003e/dev/null`\n- **Missing**: Language configuration option and auto-detection\n\n## Required Changes\n\n### 1. Update Package (`pkgs/stt-ptt/default.nix`)\nAdd support for language configuration:\n- Add `STT_LANGUAGE` environment variable (default: \"auto\" for auto-detection)\n- Modify whisper-cli call to use `-l \"$STT_LANGUAGE\"` flag\n- Support the special value \"auto\" for automatic language detection\n- Update help text to document the new environment variable\n\n### 2. Update Home Manager Module (`modules/home-manager/cli/stt-ptt.nix`)\nAdd new `language` option:\n- Add `language` option with type `types.enum [\"auto\" \"en\" \"es\" \"fr\" \"de\" \"it\" \"pt\" \"ru\" \"zh\" \"ja\" \"ko\" ...]`\n- Default should be \"auto\" (auto-detect)\n- Set `STT_LANGUAGE` environment variable in `home.sessionVariables`\n- Add documentation describing available languages and auto-detection behavior\n- Consider using `types.nullOr(types.str)` with \"auto\" as default for more flexibility\n\n### 3. Technical Details\nFrom whisper.cpp CLI documentation:\n- `-l LANG, --language LANG [en] spoken language ('auto' for auto-detect)`\n- Auto-detection analyzes audio to determine spoken language automatically\n- Specifying a language can improve accuracy if you know the language\n- Language codes follow ISO 639-1 standard (2-letter codes)\n\n## Implementation Plan\n\n### Package Changes:\n```nix\n# Add to environment variables section:\nSTT_LANGUAGE=\"${STT_LANGUAGE:-auto}\"\n\n# Modify whisper call:\n\"$WHISPER\" -m \"$STT_MODEL\" -f \"$AUDIO\" -l \"$STT_LANGUAGE\" -np -nt 2\u003e/dev/null\n\n# Update help text:\necho \" STT_LANGUAGE - Language code or 'auto' for auto-detection (default: auto)\"\n```\n\n### Module Changes:\n```nix\nlanguage = mkOption {\n type = types.enum [\"auto\" \"en\" \"es\" \"fr\" \"de\" \"it\" \"pt\" \"ru\" \"zh\" \"ja\" \"ko\"];\n default = \"auto\";\n description = ''\n Language for speech recognition. Use \"auto\" for automatic language detection,\n or specify a language code (e.g., \"en\", \"es\", \"fr\") for better accuracy.\n Auto-detection analyzes the audio to determine the spoken language.\n '';\n};\n\nhome.sessionVariables = {\n STT_LANGUAGE = cfg.language;\n # ... existing variables\n};\n```\n\n## Benefits\n- **Multilingual support**: Users can speak in any language\n- **Convenience**: Auto-detection eliminates need to specify language\n- **Accuracy**: Explicit language selection improves transcription accuracy\n- **Backward compatible**: Default behavior (auto) matches current functionality\n\n## Testing Considerations\n1. Test auto-detection with multiple languages\n2. Test explicit language selection\n3. Test fallback behavior when auto-detection fails\n4. Test with different Whisper models\n\n## Related\n- whisper.cpp supports 100+ languages\n- Language codes follow ISO 639-1 standard\n- Auto-detection adds slight processing overhead","status":"open","priority":2,"issue_type":"feature","created_at":"2026-01-10T08:40:51.478869019+01:00","created_by":"m3tam3re","updated_at":"2026-01-10T08:41:29.787702598+01:00"}
{"id":"nixpkgs-p79","title":"feat: Add auto-language detection support to stt-ptt package","description":"## Current State\n- **Package**: `/home/m3tam3re/p/NIX/nixpkgs/pkgs/stt-ptt/default.nix`\n- **Module**: `/home/m3tam3re/p/NIX/nixpkgs/modules/home-manager/cli/stt-ptt.nix`\n- **Current whisper-cli call** (line 55): `\"$WHISPER\" -m \"$STT_MODEL\" -f \"$AUDIO\" -np -nt 2\u003e/dev/null`\n- **Missing**: Language configuration option and auto-detection\n\n## Required Changes\n\n### 1. Update Package (`pkgs/stt-ptt/default.nix`)\nAdd support for language configuration:\n- Add `STT_LANGUAGE` environment variable (default: \"auto\" for auto-detection)\n- Modify whisper-cli call to use `-l \"$STT_LANGUAGE\"` flag\n- Support the special value \"auto\" for automatic language detection\n- Update help text to document the new environment variable\n\n### 2. Update Home Manager Module (`modules/home-manager/cli/stt-ptt.nix`)\nAdd new `language` option:\n- Add `language` option with type `types.enum [\"auto\" \"en\" \"es\" \"fr\" \"de\" \"it\" \"pt\" \"ru\" \"zh\" \"ja\" \"ko\" ...]`\n- Default should be \"auto\" (auto-detect)\n- Set `STT_LANGUAGE` environment variable in `home.sessionVariables`\n- Add documentation describing available languages and auto-detection behavior\n- Consider using `types.nullOr(types.str)` with \"auto\" as default for more flexibility\n\n### 3. Technical Details\nFrom whisper.cpp CLI documentation:\n- `-l LANG, --language LANG [en] spoken language ('auto' for auto-detect)`\n- Auto-detection analyzes audio to determine spoken language automatically\n- Specifying a language can improve accuracy if you know the language\n- Language codes follow ISO 639-1 standard (2-letter codes)\n\n## Implementation Plan\n\n### Package Changes:\n```nix\n# Add to environment variables section:\nSTT_LANGUAGE=\"${STT_LANGUAGE:-auto}\"\n\n# Modify whisper call:\n\"$WHISPER\" -m \"$STT_MODEL\" -f \"$AUDIO\" -l \"$STT_LANGUAGE\" -np -nt 2\u003e/dev/null\n\n# Update help text:\necho \" STT_LANGUAGE - Language code or 'auto' for auto-detection (default: auto)\"\n```\n\n### Module Changes:\n```nix\nlanguage = mkOption {\n type = types.enum [\"auto\" \"en\" \"es\" \"fr\" \"de\" \"it\" \"pt\" \"ru\" \"zh\" \"ja\" \"ko\"];\n default = \"auto\";\n description = ''\n Language for speech recognition. Use \"auto\" for automatic language detection,\n or specify a language code (e.g., \"en\", \"es\", \"fr\") for better accuracy.\n Auto-detection analyzes the audio to determine the spoken language.\n '';\n};\n\nhome.sessionVariables = {\n STT_LANGUAGE = cfg.language;\n # ... existing variables\n};\n```\n\n## Benefits\n- **Multilingual support**: Users can speak in any language\n- **Convenience**: Auto-detection eliminates need to specify language\n- **Accuracy**: Explicit language selection improves transcription accuracy\n- **Backward compatible**: Default behavior (auto) matches current functionality\n\n## Testing Considerations\n1. Test auto-detection with multiple languages\n2. Test explicit language selection\n3. Test fallback behavior when auto-detection fails\n4. Test with different Whisper models\n\n## Related\n- whisper.cpp supports 100+ languages\n- Language codes follow ISO 639-1 standard\n- Auto-detection adds slight processing overhead","status":"closed","priority":2,"issue_type":"feature","created_at":"2026-01-10T08:40:51.478869019+01:00","created_by":"m3tam3re","updated_at":"2026-01-10T09:35:49.421572334+01:00","closed_at":"2026-01-10T09:35:49.421572334+01:00","close_reason":"Implemented auto-language detection support: Added STT_LANGUAGE environment variable to package, modified whisper-cli call to use -l flag, updated help text, and added language option to HM module with enum support for 23 common languages plus 'auto'."}
{"id":"nixpkgs-sys","title":"Update config file format to include args per project","description":"Update the config file at ~/.config/rofi-project-opener/config to support per-project args. Change from PROJECT_DIRS to a format that encodes both path and args, e.g. JSON or structured text.","status":"closed","priority":1,"issue_type":"task","created_at":"2026-01-05T12:45:51.793810575+01:00","created_by":"m3tam3re","updated_at":"2026-01-05T12:48:21.75755315+01:00","closed_at":"2026-01-05T12:48:21.75755315+01:00","close_reason":"Config file now writes JSON (projects.json) with per-project path and args","dependencies":[{"issue_id":"nixpkgs-sys","depends_on_id":"nixpkgs-w3u","type":"parent-child","created_at":"2026-01-05T12:46:01.400263722+01:00","created_by":"m3tam3re"},{"issue_id":"nixpkgs-sys","depends_on_id":"nixpkgs-e2u","type":"blocks","created_at":"2026-01-05T12:46:04.055948778+01:00","created_by":"m3tam3re"}]}
{"id":"nixpkgs-w3u","title":"projectDirs should support args, for example --agent \u003cagentname\u003e","description":"This means project directories need to be an attribute set and we also need to change the launch command to \"opencode \u003cdirectory name\u003e \u003carguments\u003e\".\n","status":"closed","priority":2,"issue_type":"feature","created_at":"2026-01-05T11:56:13.844735432+01:00","created_by":"m3tam3re","updated_at":"2026-01-05T12:49:42.34144046+01:00","closed_at":"2026-01-05T12:49:42.34144046+01:00","close_reason":"Feature complete: projectDirs now supports args per directory"}