CLI reference¶

drpkit <command> [options]

  read     read Fairlight mixer settings or edit information (read-only)
  write    analysis only: write raw int32 mixer values into a new copy of a .drp
  compose  build a new .drp from a template + media + cut list

drpkit -V | --version
drpkit <command> -h

drpkit alone prints the usage and exits with 2; an unknown command also exits with 2.

drpkit read¶

drpkit read [-h] [--timeline TIMELINE] [--json] [--all] [--expect EXPECT] [--edits] path

Argument	Meaning
`path`	a `.drp` file (opened as a zip, read only) or a `Project.db` (SQLite, `mode=ro&immutable=1`). The file type is chosen by the `.drp` extension; anything else is treated as `Project.db`
`--timeline NAME`	only this timeline. Without it every timeline that has a Fairlight model is printed
`--json`	machine-readable output instead of Markdown
`--all`	also print plugin knobs that are not in the map (raw normalized value) and the raw state flags of plugins
`--expect FILE`	compare with expected values (YAML, below). Prints `一致 N 件／不一致 M 件` (matches / mismatches) and each mismatch; exit code 1 if anything differs, 0 otherwise
`--edits`	print edit information (clips, fades, keyframes, transitions, clip colors, markers) instead of the mixer

Exit codes: 0 = done; 1 = --expect mismatch; 2 = no timeline with an FLStudioModelBA found (mixer mode).

Mixer output (`--json`)¶

[
 {"timeline": "Fixture",
  "seqref": "…",
  "strips": [
   {"name": "Track1", "type": "track", "column": 0,
    "values": {"fader": {"value": "0 dB", "raw": 0, "conf": "two_point"}, "…": {}},
    "plugins": [
     {"plugin": "Voice Isolation", "slot": "内蔵FX(6)",
      "params": {"amount":  {"value": "82.01", "raw": 0.82005, "conf": "two_point"},
                 "enabled": {"value": "ON",    "raw": 0,       "conf": "direct"}}}]}]}
]

One strip per column: a stereo track gives two strips with the same name (column 0 and 1 of that table position). Buses have "type": "bus".
values keys are the item names of the map; raw is the stored int32; conf is the confidence.
slot is Slot1 … Slot6 (insert slots) or 内蔵FX(n) (track built-in FX, n = 6–10). See Plugins.
Values the reader cannot convert are reported as 未較正(raw=…) (not calibrated) or 画面空欄(raw=…) (the UI shows a blank field), with conf uncal.

`--expect` file¶

timeline: Fixture            # optional: only this timeline
strips:
  Track1: {comp_threshold: -15 dB, comp_ratio: "2.0:1", Voice Isolation.enabled: "ON"}
  Bus 1:  {limiter_on: "ON"}

Keys are item names, or Plugin.label for plugin values. Each expected value is compared as a string with the converted value exactly as printed (without the confidence mark). Examples: tests/expect_fixture_default-01.yaml, tests/expect_fixture_roundtrip-01.yaml.

Edit output (`--edits --json`)¶

[{"timeline": "Fixture", "fps": 24.0, "start_frame": 86400,
  "items": [
   {"track": "Track1", "type": "audio", "kind": "AudioClip", "name": "sine1k.wav",
    "start": 86400, "end": 86640.0, "duration": 240.0, "src_in": 0.0, "src_out": 240.0,
    "color": "Orange", "fade_in": 48.0, "fade_out": 24.0, "volume_db": -6.0,
    "volume_db_keyframes": [{"frame": 48, "value": 0.0, "flags": 786432}],
    "clip_plugins": [{"plugin": "Voice Isolation", "params": {"amount": {"value": "55.0", "raw": 0.55, "conf": "two_point"}}}],
    "start_tc": "01:00:00:00"},
   {"kind": "Transition", "transition": "Cross Fade 0DB", "alignment": "centered on edit point", "…": "…"}],
  "markers": [{"frame": 86448, "offset": 48, "color": "Red", "name": "Cut here",
               "note": "…", "duration": 1, "custom_data": ""}]}]

Key	Meaning
`start`, `end`, `duration`	timeline frames (`end` = `start + duration`)
`src_in`, `src_out`	source frames; `src_out` = `src_in + duration`
`fade_in`, `fade_out`	fade lengths in frames (only when stored)
`fade_in_curve`, `fade_out_curve`	curve point `[position %, level in 0.1 dB]` (only for curved fades)
`volume_db`	clip volume in dB
`volume_db_keyframes`	`[{frame: source frame, value: dB, flags}]`
`clip_plugins`	clip built-in FX (Voice Isolation, Dialogue Leveler); present = enabled
`start_tc`	timecode of `start` (`.drp` only)
`markers[].frame` / `offset`	absolute timeline frame / frames from the timeline start

fps and start_frame come from the .drp; for a Project.db they are null and markers carry only offset. Details: Clips & edits.

drpkit compose¶

drpkit compose [-h] --cuts CUTS --media TRACK=FILE [--media ...] [--timeline TIMELINE] [--keep-clip-fx] template output

Argument	Meaning
`template`	template `.drp` (read only)
`output`	new `.drp`; must not exist
`--cuts FILE`	cut list, `.json` or `.csv` (format). Relative `media` paths in it are resolved from the cut list's folder
`--media N=FILE`	audio file for audio track N (1-based, in template order). Every audio track of the template timeline must be given once. `N=@template` keeps the template's own media for that track (cuts, fades and keyframes are still applied)
`--timeline NAME`	template timeline; required when the template has several timelines
`--keep-clip-fx`	carry over each track's clip volume, volume keyframes and the other effects in `EffectFiltersBA` from the template's earliest clip on that track; fades and keyframes in the cut list take precedence. Without it, new clips get clip volume 0 dB. (Clip built-in FX — Voice Isolation, Dialogue Leveler — live in the clip's `FieldsBlob`, which every new clip copies from that same earliest clip in either case; see Clips & edits.)

On success it prints a JSON report (fps, timeline_start_frame, markers, cuts, media, extra_media, preservation). On any error it prints the reason and exits with 2; no output file is created.

What it changes and what it keeps: the template's audio items and their clip colors are replaced, the media-pool items of the template tracks are pointed at your files (metadata from ffprobe, see Media), the sequence extents are set to the first and last cut, and markers are written if the cut list has markers. Everything else — the Fairlight model (mixer, plugins, buses, routing), video, color, Fusion, other media-pool items — is compared before writing and must be byte-identical.

Refused rather than guessed: overlapping clips (use adjacent cuts and a crossfade), items other than audio clips and transitions on the audio tracks, the same media on several tracks, one media-pool item shared by tracks, source_out beyond the media, channel changes where the template's channel mapping is not the default, unsupported media formats.

drpkit write¶

drpkit write [-h] [--set NAME:ITEM=VALUE ...] [--dump-item ITEM] [--timeline TIMELINE] src [dst]

Argument	Meaning
`src`	input `.drp` (read only)
`dst`	output `.drp`; must not exist. Required with `--set`
`--set NAME:ITEM=VALUE`	write the int32 `VALUE` (decimal or `0x…`) into item `ITEM` (e.g. `0x10d`) of the strip `NAME` — a track name as stored in item `0xBC`, or `Bus 1`. A stereo track is written in all columns with that name. Repeatable
`--dump-item ITEM`	print that item's values in every table (first 6 columns) and exit; no output file
`--timeline NAME`	timeline to write; required when the `.drp` has several timelines with a Fairlight model

Each written cell is read back and compared; the output is created only if exactly the requested cells changed. It prints one line per write (書込 NAME 0x… = VALUE（列 …）). Use it only on copies, to build conversion tables: write a stored value → import into Resolve → read the UI.