davinci-resolve-mcp 2.155.0 → 2.156.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +27 -0
- package/README.md +3 -1
- package/README.zh-CN.md +4 -2
- package/docs/guides/native-drt-authoring.md +30 -2
- package/install.py +1 -1
- package/package.json +1 -1
- package/resolve-advanced/server/author-interchange.mjs +89 -52
- package/resolve-advanced/server/tools/editorial.mjs +1 -1
- package/src/granular/common.py +1 -1
- package/src/server.py +1 -1
package/CHANGELOG.md
CHANGED
|
@@ -2,6 +2,33 @@
|
|
|
2
2
|
|
|
3
3
|
Release history for the DaVinci Resolve MCP Server. The latest release is summarized in the root README; older entries live here to keep the README focused.
|
|
4
4
|
|
|
5
|
+
## What's New in v2.156.1 — docs catch up with the fade + QC arc
|
|
6
|
+
|
|
7
|
+
### Documentation
|
|
8
|
+
|
|
9
|
+
- `docs/guides/native-drt-authoring.md` learns the v2.148–2.156 arc: the
|
|
10
|
+
A1–A16 audio ceiling, the fade capability row and the fade/AAF-overlap
|
|
11
|
+
laws, the `blackLegs` ledger field, and the full `verify_roundtrip`
|
|
12
|
+
surface (markers, fades, retimes, audio) in the delivery checklist.
|
|
13
|
+
- The Native .drt Authoring and Headless Edit Loop guides are now linked
|
|
14
|
+
from both READMEs' guide tables — both were orphaned.
|
|
15
|
+
|
|
16
|
+
## What's New in v2.156.0 — the QC loop learns audio; the verify surface is complete
|
|
17
|
+
|
|
18
|
+
### Added
|
|
19
|
+
|
|
20
|
+
- **`editorial.verify_roundtrip` is audio-aware.** Declared audio events
|
|
21
|
+
compare pairwise through the same machinery as video — record, source,
|
|
22
|
+
retime, and fade-window excusal — with audio mismatches tagged
|
|
23
|
+
`trackType: 'audio'`, AAF channel legs deduped, and BL/silence legs merged
|
|
24
|
+
out. A video-only turnover whose re-export carries audio (the A1
|
|
25
|
+
convenience mirror this bridge authors) reports `audio.compared: false`
|
|
26
|
+
informationally instead of failing. Live loop verified: an explicit A2 leg
|
|
27
|
+
round-trips through `EXPORT_OTIO` at exact geometry.
|
|
28
|
+
- With markers (v2.147), fades (v2.153), retimes (v2.154), and now audio,
|
|
29
|
+
the round-trip QC surface covers every structure the conform bridge
|
|
30
|
+
authors.
|
|
31
|
+
|
|
5
32
|
## What's New in v2.155.0 — Premiere transitions land properly: centered spans and fades
|
|
6
33
|
|
|
7
34
|
### Fixed
|
package/README.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
English | [简体中文](README.zh-CN.md)
|
|
4
4
|
|
|
5
|
-
[](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
|
|
6
6
|
[](https://www.npmjs.com/package/davinci-resolve-mcp)
|
|
7
7
|
[](docs/reference/api-coverage.md)
|
|
8
8
|
[-blue.svg)](#server-modes)
|
|
@@ -303,6 +303,8 @@ For method-by-method status, see [API Coverage and Test Results](docs/reference/
|
|
|
303
303
|
| [Multicam Setup Helper Guide](docs/guides/multicam-setup-guide.md) | Stacked timeline prep, helper/API boundary, and Resolve UI conversion steps |
|
|
304
304
|
| [Editorial Decision Guide](docs/guides/editorial-decision-guide.md) | Project-owned editorial craft guidance for analysis and timeline decisions |
|
|
305
305
|
| [Conforming an Avid AAF](docs/guides/conforming-an-avid-aaf.md) | Why all three Resolve-native routes fail on a consolidated turnover, and which one is dangerous |
|
|
306
|
+
| [Native .drt Authoring](docs/guides/native-drt-authoring.md) | Offline template-spliced timeline authoring: cuts, retimes, transitions, fades, markers, compounds — and the measured laws behind them |
|
|
307
|
+
| [Headless Edit Loop](docs/guides/headless-edit-loop.md) | Driving Resolve from the command line: which interchange formats relink and round-trip, measured in GUI and -nogui |
|
|
306
308
|
| [Color Decision Guide](docs/guides/color-decision-guide.md) | Project-owned color correction guidance and Resolve color API boundaries |
|
|
307
309
|
| [Contributing and Project Layout](docs/contributing.md) | Contribution workflow, platform support, security notes, repository structure |
|
|
308
310
|
| [Security Policy](SECURITY.md) | Local stdio trust boundary, tool metadata, confirmation guidance, reporting |
|
package/README.zh-CN.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
[English](README.md) | 简体中文
|
|
4
4
|
|
|
5
|
-
[](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
|
|
6
6
|
[](https://www.npmjs.com/package/davinci-resolve-mcp)
|
|
7
7
|
[](docs/reference/api-coverage.md)
|
|
8
8
|
[-blue.svg)](#服务器模式)
|
|
@@ -12,7 +12,7 @@
|
|
|
12
12
|
[](https://www.python.org/downloads/)
|
|
13
13
|
[](https://opensource.org/licenses/MIT)
|
|
14
14
|
|
|
15
|
-
> 本翻译对应 v2.
|
|
15
|
+
> 本翻译对应 v2.156.1 版 README。如与英文原版有出入,以 [英文原版](README.md) 为准。
|
|
16
16
|
|
|
17
17
|
一个 Model Context Protocol (MCP) 服务器,让 AI 助手通过官方脚本 API 控制 DaVinci Resolve Studio(达芬奇)。它提供完整的 API 覆盖,外加带护栏的工作流助手,涵盖剪辑、媒体池整理、渲染设置、审阅标记、调色、Fusion、Fairlight、项目生命周期任务、扩展开发,以及不碰源媒体的媒体分析。
|
|
18
18
|
|
|
@@ -221,6 +221,8 @@ DRX 调色写入**针对 Resolve Studio 做过实机校准**:调色参数默
|
|
|
221
221
|
| [多机位设置助手指南](docs/guides/multicam-setup-guide.md) | 堆叠时间线准备、助手/API 边界、Resolve UI 转换步骤 |
|
|
222
222
|
| [剪辑决策指南](docs/guides/editorial-decision-guide.md) | 项目自有的剪辑手艺指引,用于分析与时间线决策 |
|
|
223
223
|
| [套底 Avid AAF](docs/guides/conforming-an-avid-aaf.md) | 为何三条 Resolve 原生路径在合并交接上全部失败,以及哪条最危险 |
|
|
224
|
+
| [原生 .drt 离线创作](docs/guides/native-drt-authoring.md) | 离线模板拼接式时间线创作:剪切、变速、转场、淡入淡出、标记、复合片段,以及背后的实测规律 |
|
|
225
|
+
| [无界面剪辑循环](docs/guides/headless-edit-loop.md) | 纯命令行驱动 Resolve:哪些交换格式能重新链接并往返,GUI 与 -nogui 双双实测 |
|
|
224
226
|
| [调色决策指南](docs/guides/color-decision-guide.md) | 项目自有的校色指引与 Resolve 调色 API 边界 |
|
|
225
227
|
| [贡献与项目布局](docs/contributing.md) | 贡献流程、平台支持、安全说明、仓库结构 |
|
|
226
228
|
| [安全策略](SECURITY.md) | 本地 stdio 信任边界、工具元数据、确认指引、漏洞报告 |
|
|
@@ -45,6 +45,7 @@ window. `render.verify_output` covers the container-level checks.
|
|
|
45
45
|
| Wipes (EDL `W`-codes) | `transitions[].type: 'wipe'` (single soft-edge style — host-importer parity) | v2.138 |
|
|
46
46
|
| Transition styles | `transitions[].type: 'dip' \| 'additive' \| 'fade-to-color' \| 'smooth-cut' \| 'non-additive'` (XMEML effectids route automatically) | v2.140 |
|
|
47
47
|
| Audio cross-fades | `transitions[].trackType: 'audio'` | v2.116 |
|
|
48
|
+
| Fades to/from black | EDL `BL` dissolves, OTIO gap-adjacent Transitions, XMEML/PrProj edge spans, AAF filler-adjacent Transitions — authored as Solid-Color-generator dissolves | v2.150–2.155 |
|
|
48
49
|
| Constant retimes, forward | `cuts[].speed` (e.g. `0.5`) | v2.113 |
|
|
49
50
|
| Constant retimes, reverse | `cuts[].reverse` | v2.114 |
|
|
50
51
|
| Freeze frames | `cuts[].freeze` (holds source frame `srcIn`) | v2.134 |
|
|
@@ -70,7 +71,8 @@ plus a `sourceMap` — all five formats are route-proven end-to-end
|
|
|
70
71
|
(parse → assemble → import → measured frames and RMS) — and
|
|
71
72
|
returns an honesty ledger
|
|
72
73
|
(`authoredTransitions`, `droppedTransitions` with reasons, `authoredRetimes`,
|
|
73
|
-
`flattenedRetimes`, `authoredAudioEvents`, `upperTrackCutsVideoOnly
|
|
74
|
+
`flattenedRetimes`, `authoredAudioEvents`, `upperTrackCutsVideoOnly`,
|
|
75
|
+
`blackLegs` for authored fade generators and skipped audio-silence legs).
|
|
74
76
|
|
|
75
77
|
## The laws (why the constraints are what they are)
|
|
76
78
|
|
|
@@ -153,6 +155,24 @@ for the post-cut portion. Off-center spans render (blend measured across an
|
|
|
153
155
|
uneven [cut-6, cut+18) span); everything unauthorable stays in
|
|
154
156
|
`droppedTransitions` with the reason.
|
|
155
157
|
|
|
158
|
+
**Fade law.** A transition needs a physical item on BOTH sides of its
|
|
159
|
+
junction: a hand-authored single-sided transition element (span at a lone
|
|
160
|
+
clip head or tail) refuses to import entirely, and Resolve's own EDL
|
|
161
|
+
importer silently DROPS dissolves involving the `BL` reel (the fade-in
|
|
162
|
+
vanishes; the fade-out hard-cuts into the Solid Color it creates). The
|
|
163
|
+
authorable form — used automatically by `assemble_from_interchange` for all
|
|
164
|
+
five formats — is a real clip↔Solid-Color-generator dissolve: a zero-length
|
|
165
|
+
black leg materializes/grows to cover its side of the span (empty track
|
|
166
|
+
renders black, so the growth is render-neutral), and the CMX zero-length
|
|
167
|
+
fade-in slug grows through the boundary shift. Audio fades to silence have
|
|
168
|
+
no authorable form (no silence source) and drop with the stated reason.
|
|
169
|
+
|
|
170
|
+
**AAF overlap law.** An AAF Transition CONSUMES record time: the incoming
|
|
171
|
+
component legitimately overlaps the outgoing by the transition duration.
|
|
172
|
+
The bridge trims the outgoing's tail to the overlap start and lets the
|
|
173
|
+
boundary shift re-extend it to the `CutPoint` — the AAF notional-cut
|
|
174
|
+
semantics exactly.
|
|
175
|
+
|
|
156
176
|
## Verification checklist for a delivered .drt
|
|
157
177
|
|
|
158
178
|
1. `timeline.import_timeline_checked` — expect `linked == total` for media
|
|
@@ -170,7 +190,15 @@ uneven [cut-6, cut+18) span); everything unauthorable stays in
|
|
|
170
190
|
verifier normalizes track labels, source naming, and per-source
|
|
171
191
|
TC-absolute source frames (fitting the offsets, e.g. 86400 for a
|
|
172
192
|
01:00:00:00 source) — `pass: true` means the authored timeline's export
|
|
173
|
-
matches the turnover's intent event-for-event.
|
|
193
|
+
matches the turnover's intent event-for-event. The verify surface covers
|
|
194
|
+
the whole authored family: timeline markers (`markersNotInExport` flags a
|
|
195
|
+
marker-less export instead of failing — EXPORT_OTIO drops them), fades
|
|
196
|
+
(BL/Solid-Color legs merge out as `blackSegments`; fade-window boundary
|
|
197
|
+
reshapes are excused into `fadeReshapedBoundaries`, never silently),
|
|
198
|
+
retimes (speed/reverse compare pairwise — EXPORT_OTIO round-trips an
|
|
199
|
+
authored Sm2TimeMap as LinearTimeWarp), and audio (declared audio events
|
|
200
|
+
compare with `trackType: 'audio'` tags; the A1 convenience mirror on a
|
|
201
|
+
video-only turnover stays informational).
|
|
174
202
|
|
|
175
203
|
## References
|
|
176
204
|
|
package/install.py
CHANGED
|
@@ -37,7 +37,7 @@ from src.utils.update_check import (
|
|
|
37
37
|
|
|
38
38
|
# ─── Version ──────────────────────────────────────────────────────────────────
|
|
39
39
|
|
|
40
|
-
VERSION = "2.
|
|
40
|
+
VERSION = "2.156.1"
|
|
41
41
|
# Only hard floor: mcp[cli] requires Python 3.10+. There is no upper bound —
|
|
42
42
|
# Resolve's scripting bridge loads into newer interpreters on recent builds
|
|
43
43
|
# (Python 3.14 verified against Resolve Studio 20.3.2). Older Resolve builds
|
package/package.json
CHANGED
|
@@ -838,16 +838,21 @@ export function verifyRoundtrip(inputEvents, exportedEvents, opts = {}) {
|
|
|
838
838
|
// before the D event — a pairing placeholder, never a rendered clip, and
|
|
839
839
|
// no export reproduces it.
|
|
840
840
|
const vids = (evts) => evts.filter((e) => /^V\d*$/.test(String(e.track)) && e.recIn != null && e.recOut != null && e.recOut > e.recIn);
|
|
841
|
-
const
|
|
841
|
+
const audsOf = (evts) => evts.filter((e) => /^A\d*$/.test(String(e.track)) && e.recIn != null && e.recOut != null && e.recOut > e.recIn);
|
|
842
|
+
const anchorOfSide = (evts) => {
|
|
842
843
|
const v = vids(evts);
|
|
844
|
+
if (v.length) return Math.min(...v.map((e) => e.recIn));
|
|
845
|
+
const au = audsOf(evts);
|
|
846
|
+
return au.length ? Math.min(...au.map((e) => e.recIn)) : 0;
|
|
847
|
+
};
|
|
848
|
+
const norm = (evts, pick = vids) => {
|
|
849
|
+
const v = pick(evts);
|
|
843
850
|
if (!v.length) return [];
|
|
844
|
-
const off =
|
|
851
|
+
const off = anchorOfSide(evts);
|
|
845
852
|
return v
|
|
846
853
|
.map((e) => ({ track: canonTrack(e.track), source: mapSource(canonSource(e.source)), recIn: e.recIn - off, recOut: e.recOut - off, srcIn: e.srcIn ?? 0, speed: e.speed ?? 100, reverse: Boolean(e.reverse) }))
|
|
847
854
|
.sort((a, b) => a.track.localeCompare(b.track) || a.recIn - b.recIn);
|
|
848
855
|
};
|
|
849
|
-
const a0 = norm(inputEvents);
|
|
850
|
-
const b0 = norm(exportedEvents);
|
|
851
856
|
// FADES (E94): BL legs on the input side and the Solid Color generators a
|
|
852
857
|
// fade conform authors for them are the same thing — BLACK. Black segments
|
|
853
858
|
// are synthesized filler whose extents follow from the picture boundaries,
|
|
@@ -857,60 +862,91 @@ export function verifyRoundtrip(inputEvents, exportedEvents, opts = {}) {
|
|
|
857
862
|
// start-at-cut reshaping), so an edge within an input junction's fade
|
|
858
863
|
// window is excused — and reported, not silently absorbed.
|
|
859
864
|
const isBlackSeg = (e) => /^(bl|black|solid color)$/.test(e.source);
|
|
860
|
-
const
|
|
861
|
-
const
|
|
862
|
-
|
|
863
|
-
|
|
864
|
-
const inOff = rawVids.length ? Math.min(...rawVids.map((e) => e.recIn)) : 0;
|
|
865
|
-
const junctions = inputEvents
|
|
866
|
-
.filter((e) => /^V\d*$/.test(String(e.track)) && e.recIn != null && e.transition && e.transition.duration > 0)
|
|
867
|
-
.map((e) => ({ frame: e.recIn - inOff, d: e.transition.duration }));
|
|
868
|
-
const inFadeWindow = (edgeIn, edgeExp) => junctions.some(
|
|
869
|
-
(j) => Math.abs(edgeIn - j.frame) <= j.d && Math.abs(edgeExp - edgeIn) <= j.d,
|
|
870
|
-
);
|
|
865
|
+
const inAnchor = anchorOfSide(inputEvents);
|
|
866
|
+
const junctionsFor = (trackRe) => inputEvents
|
|
867
|
+
.filter((e) => trackRe.test(String(e.track)) && e.recIn != null && e.transition && e.transition.duration > 0)
|
|
868
|
+
.map((e) => ({ frame: e.recIn - inAnchor, d: e.transition.duration }));
|
|
871
869
|
const fadeReshapedBoundaries = [];
|
|
872
870
|
const mismatches = [];
|
|
873
|
-
if (a.length !== b.length) mismatches.push({ kind: 'count', input: a.length, exported: b.length });
|
|
874
871
|
const srcOffsets = {};
|
|
875
|
-
const
|
|
876
|
-
|
|
877
|
-
const
|
|
878
|
-
|
|
879
|
-
|
|
880
|
-
|
|
881
|
-
const
|
|
882
|
-
|
|
883
|
-
const
|
|
884
|
-
|
|
885
|
-
if (
|
|
886
|
-
|
|
887
|
-
|
|
888
|
-
|
|
872
|
+
const comparePairs = (a, b, junctions, trackType) => {
|
|
873
|
+
const tt = trackType === 'audio' ? { trackType: 'audio' } : {};
|
|
874
|
+
const inFadeWindow = (edgeIn, edgeExp) => junctions.some(
|
|
875
|
+
(j) => Math.abs(edgeIn - j.frame) <= j.d && Math.abs(edgeExp - edgeIn) <= j.d,
|
|
876
|
+
);
|
|
877
|
+
if (a.length !== b.length) mismatches.push({ kind: 'count', ...tt, input: a.length, exported: b.length });
|
|
878
|
+
const n = Math.min(a.length, b.length);
|
|
879
|
+
for (let i = 0; i < n; i += 1) {
|
|
880
|
+
const x = a[i], y = b[i];
|
|
881
|
+
if (x.track !== y.track) { mismatches.push({ kind: 'track', ...tt, at: i, input: x.track, exported: y.track }); continue; }
|
|
882
|
+
if (x.source !== y.source) { mismatches.push({ kind: 'source', ...tt, at: i, input: x.source, exported: y.source }); continue; }
|
|
883
|
+
const inBad = Math.abs(x.recIn - y.recIn) > recTol;
|
|
884
|
+
const outBad = Math.abs(x.recOut - y.recOut) > recTol;
|
|
885
|
+
if (inBad || outBad) {
|
|
886
|
+
const inExcused = !inBad || inFadeWindow(x.recIn, y.recIn);
|
|
887
|
+
const outExcused = !outBad || inFadeWindow(x.recOut, y.recOut);
|
|
888
|
+
if (inExcused && outExcused) {
|
|
889
|
+
fadeReshapedBoundaries.push({ at: i, ...tt, source: x.source, input: [x.recIn, x.recOut], exported: [y.recIn, y.recOut] });
|
|
890
|
+
} else {
|
|
891
|
+
mismatches.push({ kind: 'record', ...tt, at: i, input: [x.recIn, x.recOut], exported: [y.recIn, y.recOut] });
|
|
892
|
+
continue;
|
|
893
|
+
}
|
|
894
|
+
}
|
|
895
|
+
// RETIMES (E95): a conform that lost its retime is a wrong timeline
|
|
896
|
+
// that record/source geometry alone cannot catch (the record extent is
|
|
897
|
+
// unchanged; only the playback rate is). Resolve's EXPORT_OTIO carries
|
|
898
|
+
// an authored Sm2TimeMap back as LinearTimeWarp (measured live: 50%
|
|
899
|
+
// in → time_scalar 0.5 out), so a speed/reverse mismatch is real drift.
|
|
900
|
+
if (Math.abs((x.speed ?? 100) - (y.speed ?? 100)) > 0.5 || x.reverse !== y.reverse) {
|
|
901
|
+
mismatches.push({
|
|
902
|
+
kind: 'retime', ...tt, at: i, source: x.source,
|
|
903
|
+
input: { speed: x.speed, reverse: x.reverse },
|
|
904
|
+
exported: { speed: y.speed, reverse: y.reverse },
|
|
905
|
+
});
|
|
889
906
|
continue;
|
|
890
907
|
}
|
|
908
|
+
// A fade-reshaped head trims record AND source together (source stays
|
|
909
|
+
// record-aligned), so the per-source constant offset is fitted net of
|
|
910
|
+
// the record shift — otherwise a source cut both plain and faded would
|
|
911
|
+
// read as a source-frames drift.
|
|
912
|
+
const off = (y.srcIn - (y.recIn - x.recIn)) - x.srcIn;
|
|
913
|
+
if (srcOffsets[x.source] === undefined) srcOffsets[x.source] = off;
|
|
914
|
+
else if (Math.abs(off - srcOffsets[x.source]) > srcTol) {
|
|
915
|
+
mismatches.push({ kind: 'source-frames', ...tt, at: i, source: x.source, expectedOffset: srcOffsets[x.source], gotOffset: off });
|
|
916
|
+
}
|
|
891
917
|
}
|
|
892
|
-
|
|
893
|
-
|
|
894
|
-
|
|
895
|
-
|
|
896
|
-
|
|
897
|
-
|
|
898
|
-
|
|
899
|
-
|
|
900
|
-
|
|
901
|
-
|
|
902
|
-
|
|
903
|
-
|
|
904
|
-
|
|
905
|
-
|
|
906
|
-
|
|
907
|
-
|
|
908
|
-
|
|
909
|
-
|
|
910
|
-
|
|
911
|
-
|
|
912
|
-
|
|
913
|
-
|
|
918
|
+
return n;
|
|
919
|
+
};
|
|
920
|
+
|
|
921
|
+
const a0 = norm(inputEvents);
|
|
922
|
+
const b0 = norm(exportedEvents);
|
|
923
|
+
const a = a0.filter((e) => !isBlackSeg(e));
|
|
924
|
+
const b = b0.filter((e) => !isBlackSeg(e));
|
|
925
|
+
const blackSegments = { input: a0.length - a.length, exported: b0.length - b.length };
|
|
926
|
+
const n = comparePairs(a, b, junctionsFor(/^V\d*$/), 'video');
|
|
927
|
+
|
|
928
|
+
// AUDIO (E97): compared only when the INPUT declares audio events — a
|
|
929
|
+
// video-only turnover legitimately re-exports with audio (the A1
|
|
930
|
+
// convenience mirror), which is reported informationally, never failed.
|
|
931
|
+
// AAF channel legs dedupe (same track/source/range arrives once per
|
|
932
|
+
// channel); BL/silence legs merge out like video black.
|
|
933
|
+
const dedupeAud = (list) => {
|
|
934
|
+
const seen = new Set();
|
|
935
|
+
return list.filter((e) => {
|
|
936
|
+
const k = `${e.track}|${e.source}|${e.recIn}|${e.recOut}|${e.srcIn}`;
|
|
937
|
+
if (seen.has(k)) return false;
|
|
938
|
+
seen.add(k);
|
|
939
|
+
return true;
|
|
940
|
+
});
|
|
941
|
+
};
|
|
942
|
+
const aa = dedupeAud(norm(inputEvents, audsOf).filter((e) => !isBlackSeg(e)));
|
|
943
|
+
const ba = dedupeAud(norm(exportedEvents, audsOf).filter((e) => !isBlackSeg(e)));
|
|
944
|
+
let audio;
|
|
945
|
+
if (aa.length) {
|
|
946
|
+
audio = { input: aa.length, exported: ba.length, compared: true };
|
|
947
|
+
comparePairs(aa, ba, junctionsFor(/^A\d*$/), 'audio');
|
|
948
|
+
} else if (ba.length) {
|
|
949
|
+
audio = { input: 0, exported: ba.length, compared: false, note: 'export-only audio (A1 convenience mirror) — not compared' };
|
|
914
950
|
}
|
|
915
951
|
// MARKERS (E88): compare track-'MARKER' pseudo-events, min-anchored to
|
|
916
952
|
// each side's own VIDEO record origin like everything else. When the
|
|
@@ -948,5 +984,6 @@ export function verifyRoundtrip(inputEvents, exportedEvents, opts = {}) {
|
|
|
948
984
|
...(markersNotInExport ? { markersNotInExport } : {}),
|
|
949
985
|
...(blackSegments.input || blackSegments.exported ? { blackSegments } : {}),
|
|
950
986
|
...(fadeReshapedBoundaries.length ? { fadeReshapedBoundaries } : {}),
|
|
987
|
+
...(audio ? { audio } : {}),
|
|
951
988
|
};
|
|
952
989
|
}
|
|
@@ -86,7 +86,7 @@ const markerSchema = z.object({
|
|
|
86
86
|
export const editorialTool = {
|
|
87
87
|
name: 'editorial',
|
|
88
88
|
description:
|
|
89
|
-
'Editorial integrity (Cluster E) — turnover interchange → normalized events → changelist + conform manifest with TIMING silent-lie guards (flattened retime / dropped J/L-cut audio / framerate-pulldown slip / reverse dropped / transition-handle starvation → flag, skip-not-fake). Report-only (gate: review). Actions: parse_interchange (EDL/OTIO/XMEML natively + AAF via pyaaf2 + PRPROJ via gunzip+XML → normalized events incl. span-explicit transitions and BL fade legs; for AAF/PRPROJ pass the file PATH as content; AAF also returns per-sequence startTimecode/startFrame — build the timeline at THAT start, not the Resolve 01:00:00:00 default — and per-clip `geometry` for Avid transform effects), list_sequences (ONE offline picker entry point across xml/edl/otio/drt/drp/aaf/prproj → [{id,name,eventCount}], plus startTimecode/startFrame for AAF), convert_to_interchange (author OTIO/EDL/DRT Resolve CAN import from events or a parsed source — the .prproj→Resolve conform bridge, no Premiere needed; editorial timing/transitions survive and per-clip effects/color do not. SPEED/REVERSE survive on the otio (LinearTimeWarp) and edl (M2) targets ONLY — this FLAT drt target flattens every retime to 100% forward and returns `flattened`/`flattenedCount` naming each event that lost one (`flattened` is always present on `drt`, empty when there were none); for a .drt that AUTHORS retimes/dissolves/multi-track/audio, use drt.assemble_from_interchange), turnover_changelist (diff old vs new → moved/retimed/replaced/new/gone + timing flags), conform_manifest (per-event assert: source resolved/handles/retime/reverse/TC-base), marker_roundtrip (markers with provenance tags), verify_roundtrip (input events vs re-export events -> pass/mismatches + fitted per-source TC offsets + marker compare w/ markersNotInExport honesty flag; FADE-AWARE: BL/Solid-Color legs merge out as blackSegments and fade-window boundary reshapes are excused into fadeReshapedBoundaries instead of failing; RETIME-AWARE: speed/reverse compare pairwise — EXPORT_OTIO carries an authored Sm2TimeMap back as LinearTimeWarp (measured), so a flattened/lost retime fails as drift geometry alone cannot catch; the conform QC loop-closer). Offline (AAF needs pyaaf2; live AAF/DRP import is on the Python davinci-resolve MCP).',
|
|
89
|
+
'Editorial integrity (Cluster E) — turnover interchange → normalized events → changelist + conform manifest with TIMING silent-lie guards (flattened retime / dropped J/L-cut audio / framerate-pulldown slip / reverse dropped / transition-handle starvation → flag, skip-not-fake). Report-only (gate: review). Actions: parse_interchange (EDL/OTIO/XMEML natively + AAF via pyaaf2 + PRPROJ via gunzip+XML → normalized events incl. span-explicit transitions and BL fade legs; for AAF/PRPROJ pass the file PATH as content; AAF also returns per-sequence startTimecode/startFrame — build the timeline at THAT start, not the Resolve 01:00:00:00 default — and per-clip `geometry` for Avid transform effects), list_sequences (ONE offline picker entry point across xml/edl/otio/drt/drp/aaf/prproj → [{id,name,eventCount}], plus startTimecode/startFrame for AAF), convert_to_interchange (author OTIO/EDL/DRT Resolve CAN import from events or a parsed source — the .prproj→Resolve conform bridge, no Premiere needed; editorial timing/transitions survive and per-clip effects/color do not. SPEED/REVERSE survive on the otio (LinearTimeWarp) and edl (M2) targets ONLY — this FLAT drt target flattens every retime to 100% forward and returns `flattened`/`flattenedCount` naming each event that lost one (`flattened` is always present on `drt`, empty when there were none); for a .drt that AUTHORS retimes/dissolves/multi-track/audio, use drt.assemble_from_interchange), turnover_changelist (diff old vs new → moved/retimed/replaced/new/gone + timing flags), conform_manifest (per-event assert: source resolved/handles/retime/reverse/TC-base), marker_roundtrip (markers with provenance tags), verify_roundtrip (input events vs re-export events -> pass/mismatches + fitted per-source TC offsets + marker compare w/ markersNotInExport honesty flag; FADE-AWARE: BL/Solid-Color legs merge out as blackSegments and fade-window boundary reshapes are excused into fadeReshapedBoundaries instead of failing; RETIME-AWARE: speed/reverse compare pairwise — EXPORT_OTIO carries an authored Sm2TimeMap back as LinearTimeWarp (measured), so a flattened/lost retime fails as drift geometry alone cannot catch; AUDIO-AWARE: declared audio events compare (channel legs deduped, mismatches tagged trackType audio) while the mirrored-A1 export of a video-only turnover stays informational; the conform QC loop-closer). Offline (AAF needs pyaaf2; live AAF/DRP import is on the Python davinci-resolve MCP).',
|
|
90
90
|
async handler({ action, args }) {
|
|
91
91
|
if (action === 'parse_interchange') {
|
|
92
92
|
const p = parseSchema.parse(args);
|
package/src/granular/common.py
CHANGED
|
@@ -87,7 +87,7 @@ if not logging.getLogger().handlers:
|
|
|
87
87
|
handlers=[logging.StreamHandler()],
|
|
88
88
|
)
|
|
89
89
|
|
|
90
|
-
VERSION = "2.
|
|
90
|
+
VERSION = "2.156.1"
|
|
91
91
|
logger = logging.getLogger("davinci-resolve-mcp")
|
|
92
92
|
logger.info(f"Starting DaVinci Resolve MCP Server v{VERSION}")
|
|
93
93
|
logger.info(f"Detected platform: {get_platform()}")
|