Tasch/whisper-transcribe
1
0
Fork 0
Code Issues 4 Pull requests Projects Releases Packages Wiki Activity Actions 11

feat: add StructureService for LLM-powered transcript structuring #9

Merged
SaschaFuksa merged 2 commits from feat/transcript-structuring into master 2026-04-18 19:08:30 +00:00
Conversation 0 Commits 2 Files changed 8 +472 -115
Tasch commented 2026-04-18 18:52:38 +00:00
Owner
Copy link

Overview

Adds Ollama-powered service to transform raw Whisper transcripts into structured Markdown.

What's new

  • StructureService (src/structuring/structure_service.py): Calls Ollama API to structure raw transcripts
  • System Prompt (src/prompts/structure_transcript.system): Defines structuring rules (H1/H2/H3, filler removal, typo correction)
  • Integration Tests (tests/test_structure_service.py): Tests for local execution with Ollama
  • Config (settings.yaml): New llm section with ollama_url, ollama_model, prompt_path

Architecture

Whisper (raw text) → StructureService (Ollama) → Structured MD → RepoOutput / TaschOutput

Config

llm:
  ollama_url: "http://192.168.178.50:11434"
  ollama_model: "ministral-3:3b"
  prompt_path: "src/prompts/structure_transcript.system"

Running tests

Requires Ollama running at configured URL with model loaded:

python -m pytest tests/test_structure_service.py -v
## Overview Adds Ollama-powered service to transform raw Whisper transcripts into structured Markdown. ## What's new - **StructureService** (`src/structuring/structure_service.py`): Calls Ollama API to structure raw transcripts - **System Prompt** (`src/prompts/structure_transcript.system`): Defines structuring rules (H1/H2/H3, filler removal, typo correction) - **Integration Tests** (`tests/test_structure_service.py`): Tests for local execution with Ollama - **Config** (`settings.yaml`): New `llm` section with `ollama_url`, `ollama_model`, `prompt_path` ## Architecture ``` Whisper (raw text) → StructureService (Ollama) → Structured MD → RepoOutput / TaschOutput ``` ## Config ```yaml llm: ollama_url: "http://192.168.178.50:11434" ollama_model: "ministral-3:3b" prompt_path: "src/prompts/structure_transcript.system" ``` ## Running tests Requires Ollama running at configured URL with model loaded: ```bash python -m pytest tests/test_structure_service.py -v ```
feat: add StructureService for LLM-powered transcript structuring
Some checks are pending
CI / lint (pull_request) Waiting to run
Details
CI / test (pull_request) Waiting to run
Details
a06145f2aa 
Adds Ollama-powered service to transform raw Whisper transcripts into
structured Markdown with proper headings, filler word removal, and
typo correction.

New files:
- src/structuring/structure_service.py: StructureService with Ollama integration
- src/prompts/structure_transcript.system: System prompt for structuring
- tests/test_structure_service.py: Integration tests (run locally with Ollama)

Config changes:
- settings.yaml: llm section with ollama_url, ollama_model, prompt_path
fix: address review feedback
Some checks are pending
CI / lint (pull_request) Waiting to run
Details
CI / test (pull_request) Waiting to run
Details
a5f1cf9d55 
- Prompts now use Jinja2 (.j2) templates in English
- Removed prompt_path from config (prompts not user-editable)
- Added Pydantic models (LLMConfig) for config validation
- Updated system prompt to English only
- Added AGENTS.md with project conventions
- Updated tests to match new behavior
SaschaFuksa deleted branch feat/transcript-structuring 2026-04-18 19:08:30 +00:00
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels
No items
No labels
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
Tasch/whisper-transcribe!9
No description provided.