formula/analyze-single

Formula Extraction — analyze reference videos and extract reusable style DNA.

Category: formula
Source: workflows/formula/extract.py

Input Schema

Field	Type	Default	Description
formula	object	—
formula_name	string	"untitled"
media_path	object	—
regenerate	object	—	When set, this run is a regeneration. Workflows may read direction / keep / extra_instructions to modulate prompts; the engine persists parent_run_id and parent_variant_index as run lineage columns.
variants	integer	1	Number of independent variant executions (1–10). When > 1, the engine runs the workflow N times with different sampling, producing N outputs.
video_path	object	—
video_paths	any[]	—

Output Schema

No schema defined.

Task Pipeline

probe_media → transcribe_for_analysis → extract_keyframes → analyze_visuals → analyze_audio → analyze_script → analyze_effects → merge_analyses

Task	Description
probe_media	Use ffprobe to extract technical specs from the video.
transcribe_for_analysis	Transcribe the reference video. Wraps the existing Whisper pipeline.
extract_keyframes	Extract representative frames using ffmpeg scene detection + uniform sampling.
analyze_visuals	Analyze visual style from keyframes using Gemini vision.
analyze_audio	Analyze audio characteristics from transcript and probe data.
analyze_script	Analyze script structure, hook type, and narration patterns.
analyze_effects	Analyze visual effects, subtitles, overlays, and color grading from keyframes.
merge_analyses	Merge results from the 4 parallel analysis branches.

Run-spec example

Save the YAML below as my-run.yaml, edit the values, and run with the CLI or POST it to the API. Required fields are uncommented; optional knobs are documented above the input: block — copy any line under input: and uncomment to set.

workflow: formula/analyze-single

# Optional fields — copy any line(s) under `input:` and uncomment to set:
#   formula: null
#
#   formula_name: untitled
#
#   media_path: null
#
#   video_path: null
#
#   video_paths: []
#

input: {}

Run it locally:

fab-workflow --from-file my-run.yaml

Or submit over the wire — the same file is the request body:

curl -X POST 'https://gofabric.dev/v1/workflows/runs?name=formula/analyze-single' \
  -H 'Authorization: Bearer fab_xxx' \
  -H 'content-type: application/yaml' \
  --data-binary @my-run.yaml

Every workflow also accepts the universal WorkflowInput fields — variants (1–10 fan-out) and regenerate (creative-direction hints with run lineage). See Run-specs (YAML / TOML / JSON) for the full top-level shape (metadata, priority, bundle, parent, etc.).

Warnings

• Last user task merge_analyses has no Pydantic return type — workflow output schema is null. Declare a WorkflowOutput subclass and pass it to Flow(output=…) for a strict contract.
• Task merge_analyses has no Pydantic types — contract is opaque to consumers.