davinci-resolve-mcp 2.153.0 → 2.155.0

package/CHANGELOG.md

@@ -2,6 +2,43 @@
 
 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.
 
+## What's New in v2.155.0 — Premiere transitions land properly: centered spans and fades
+
+### Fixed
+
+- **Every centered Premiere transition was silently dropped**: the .prproj
+  reader attached transitions only to clips starting exactly at the span
+  start — the CMX start-at-cut shape — so a transition centered on its cut
+  (Premiere's default) matched nothing and vanished. The incoming clip is
+  now found anywhere inside the span, and `recStart` carries Premiere's
+  explicit record span to the bridge so the editor's actual alignment is
+  reproduced.
+
+### Added
+
+- **.prproj fades**: an edge span with a missing neighbor synthesizes BL
+  legs through the same black machinery as EDL/OTIO/XMEML/AAF — all five
+  ingest formats now author fades. Audio transition candidates honor
+  `recStart` too. The emitted spec shapes are the render-verified classes
+  from v2.150–152.
+
+## What's New in v2.154.0 — the QC loop learns retimes
+
+### Added
+
+- **`editorial.verify_roundtrip` is retime-aware.** A conform that lost its
+  retime is a wrong timeline that record/source geometry alone cannot catch
+  — the record extent is unchanged, only the playback rate. Speed and
+  reverse now compare pairwise (`kind: 'retime'` on drift).
+
+### Measured (good news, for once)
+
+- **`EXPORT_OTIO` carries an authored `Sm2TimeMap` back as
+  `LinearTimeWarp`**: a 50% M2 retime conformed offline reads back out of
+  Resolve's OTIO export as `time_scalar: 0.5`. The retime loop closes
+  losslessly through OTIO — live loop verified both ways (pass on the true
+  export, `retime` drift on a speed-stripped one).
+
 ## What's New in v2.153.0 — the QC loop learns fades
 
 ### Added

package/README.md

@@ -2,7 +2,7 @@
 
 English | [简体中文](README.zh-CN.md)
 
-[![Version](https://img.shields.io/badge/version-2.153.0-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
+[![Version](https://img.shields.io/badge/version-2.155.0-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
 [![npm](https://img.shields.io/npm/v/davinci-resolve-mcp.svg?label=npm&color=CB3837)](https://www.npmjs.com/package/davinci-resolve-mcp)
 [![API Coverage](https://img.shields.io/badge/API%20Coverage-100%25-brightgreen.svg)](docs/reference/api-coverage.md)
 [![Tools](https://img.shields.io/badge/MCP%20Tools-36%20(353%20full)-blue.svg)](#server-modes)

package/README.zh-CN.md

@@ -2,7 +2,7 @@
 
 [English](README.md) | 简体中文
 
-[![Version](https://img.shields.io/badge/version-2.153.0-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
+[![Version](https://img.shields.io/badge/version-2.155.0-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
 [![npm](https://img.shields.io/npm/v/davinci-resolve-mcp.svg?label=npm&color=CB3837)](https://www.npmjs.com/package/davinci-resolve-mcp)
 [![API Coverage](https://img.shields.io/badge/API%20Coverage-100%25-brightgreen.svg)](docs/reference/api-coverage.md)
 [![Tools](https://img.shields.io/badge/MCP%20Tools-36%20(353%20full)-blue.svg)](#服务器模式)
@@ -12,7 +12,7 @@
 [![Python](https://img.shields.io/badge/python-3.10+-green.svg)](https://www.python.org/downloads/)
 [![License](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 
-> 本翻译对应 v2.153.0 版 README。如与英文原版有出入，以 [英文原版](README.md) 为准。
+> 本翻译对应 v2.155.0 版 README。如与英文原版有出入，以 [英文原版](README.md) 为准。
 
 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。
 

package/install.py

@@ -37,7 +37,7 @@ from src.utils.update_check import (
 
 # ─── Version ──────────────────────────────────────────────────────────────────
 
-VERSION = "2.153.0"
+VERSION = "2.155.0"
 # Only hard floor: mcp[cli] requires Python 3.10+. There is no upper bound —
 # Resolve's scripting bridge loads into newer interpreters on recent builds
 # (Python 3.14 verified against Resolve Studio 20.3.2). Older Resolve builds

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "davinci-resolve-mcp",
-  "version": "2.153.0",
+  "version": "2.155.0",
   "description": "NPM bootstrapper for the DaVinci Resolve MCP Server.",
   "license": "MIT",
   "author": "Samuel Gursky <samgursky@gmail.com>",

package/resolve-advanced/server/author-interchange.mjs

@@ -432,9 +432,13 @@ export function eventsToAssembleSpec(events, opts = {}) {
       let pre = Math.floor(d / 2);
       if (e.transition.alignment === 'start') pre = 0;
       else if (e.transition.inOffset != null) pre = Math.min(d, Math.max(0, toTl(e.transition.inOffset, e.fps)));
+      else if (e.transition.recStart != null) {
+        const spanStartAbs = ORIGIN + (toTl(e.transition.recStart, e.fps) - minRec);
+        pre = Math.min(d, Math.max(0, recIn - spanStartAbs));
+      }
       audioTransCandidates.push({
         atFrame: recIn, durationFrames: d, track, pre,
-        explicitSpan: e.transition.alignment != null || e.transition.inOffset != null,
+        explicitSpan: e.transition.alignment != null || e.transition.inOffset != null || e.transition.recStart != null,
         index: e.index, type: e.transition.type, rawDuration: e.transition.duration,
         source: e.source, srcIn: cut.srcIn, incomingCutRef: cut, cutPoint: e.transition.cutPoint,
       });
@@ -839,7 +843,7 @@ export function verifyRoundtrip(inputEvents, exportedEvents, opts = {}) {
     if (!v.length) return [];
     const off = Math.min(...v.map((e) => e.recIn));
     return v
-      .map((e) => ({ track: canonTrack(e.track), source: mapSource(canonSource(e.source)), recIn: e.recIn - off, recOut: e.recOut - off, srcIn: e.srcIn ?? 0 }))
+      .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) }))
       .sort((a, b) => a.track.localeCompare(b.track) || a.recIn - b.recIn);
   };
   const a0 = norm(inputEvents);
@@ -885,6 +889,19 @@ export function verifyRoundtrip(inputEvents, exportedEvents, opts = {}) {
         continue;
       }
     }
+    // RETIMES (E95): a conform that lost its retime is a wrong timeline that
+    // record/source geometry alone cannot catch (the record extent is
+    // unchanged; only the playback rate is). Resolve's EXPORT_OTIO carries
+    // an authored Sm2TimeMap back as LinearTimeWarp (measured live: 50%
+    // in → time_scalar 0.5 out), so a speed/reverse mismatch is real drift.
+    if (Math.abs((x.speed ?? 100) - (y.speed ?? 100)) > 0.5 || x.reverse !== y.reverse) {
+      mismatches.push({
+        kind: 'retime', at: i, source: x.source,
+        input: { speed: x.speed, reverse: x.reverse },
+        exported: { speed: y.speed, reverse: y.reverse },
+      });
+      continue;
+    }
     // A fade-reshaped head trims record AND source together (source stays
     // record-aligned), so the per-source constant offset is fitted net of
     // the record shift — otherwise a source cut both plain and faded would

package/resolve-advanced/server/prproj.mjs

@@ -8,7 +8,7 @@
  *
  * Time is in TICKS (254,016,000,000 per second — factors by every standard frame/sample rate).
  * We derive from tick geometry alone: cuts, source in/out, timeline position, SPEED/retime
- * (srcDur/recDur), REVERSE (in>out), transitions (VideoTransitionTrackItem span), and markers.
+ * (srcDur/recDur), REVERSE (in>out), transitions (VideoTransitionTrackItem span, recStart-explicit; edge spans synthesize BL fade legs), and markers.
  *
  * HONEST limits (skip-not-fake): the schema is proprietary and version-drifting (Project Version
  * 25→42+); editorial timing/structure decodes with high fidelity, but per-clip EFFECTS / Lumetri
@@ -179,10 +179,29 @@ function walkSequence(seqEntry, byId) {
           transitions.push({ recIn: ticksToFrames(childText(cEntry.node, 'Start'), fps), duration: dur || 0 });
         }
       }
-      // Attach each transition to the clip that begins at its record position (best-effort).
+      // Attach each transition to the INCOMING clip whose record-in falls
+      // inside its span — Premiere stores the explicit record span, so
+      // recStart carries it and the bridge reproduces the editor's actual
+      // alignment (the old exact-start match silently dropped every
+      // centered transition). A span with no outgoing clip is a fade-IN,
+      // one with no incoming a fade-OUT — both synthesize BL legs for the
+      // bridge's black machinery (E91-E96).
       for (const tr of transitions) {
-        const hit = clipEvents.find((e) => e.recIn === tr.recIn);
-        if (hit) hit.transition = { type: 'dissolve', duration: tr.duration };
+        const end = tr.recIn + tr.duration;
+        const trans = { type: 'dissolve', duration: tr.duration, recStart: tr.recIn };
+        const incoming = clipEvents.find((e) => e.recIn > tr.recIn - 1 && e.recIn <= end);
+        if (incoming) {
+          incoming.transition = trans;
+          const outgoing = clipEvents.find((e) => e !== incoming && e.recOut > tr.recIn - 1 && e.recOut <= end);
+          if (!outgoing) {
+            clipEvents.push({ index: idx++, track: trackKind, source: 'BL', srcIn: 0, srcOut: 0, recIn: incoming.recIn, recOut: incoming.recIn, speed: 100, reverse: false, transition: null, fps });
+          }
+        } else {
+          const outgoing = clipEvents.find((e) => e.recOut > tr.recIn - 1 && e.recOut <= end);
+          if (outgoing) {
+            clipEvents.push({ index: idx++, track: trackKind, source: 'BL', srcIn: 0, srcOut: 0, recIn: outgoing.recOut, recOut: outgoing.recOut, speed: 100, reverse: false, transition: trans, fps });
+          }
+        }
       }
       events.push(...clipEvents);
     }

package/resolve-advanced/server/tools/editorial.mjs

@@ -86,7 +86,7 @@ const markerSchema = z.object({
 export const editorialTool = {
   name: 'editorial',
   description:
-    '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; 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; the conform QC loop-closer). Offline (AAF needs pyaaf2; live AAF/DRP import is on the Python davinci-resolve MCP).',
+    '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).',
   async handler({ action, args }) {
     if (action === 'parse_interchange') {
       const p = parseSchema.parse(args);

package/src/granular/common.py

@@ -87,7 +87,7 @@ if not logging.getLogger().handlers:
         handlers=[logging.StreamHandler()],
     )
 
-VERSION = "2.153.0"
+VERSION = "2.155.0"
 logger = logging.getLogger("davinci-resolve-mcp")
 logger.info(f"Starting DaVinci Resolve MCP Server v{VERSION}")
 logger.info(f"Detected platform: {get_platform()}")

package/src/server.py

@@ -11,7 +11,7 @@ Usage:
     python src/server.py --full       # Start the 353-tool granular server instead
 """
 
-VERSION = "2.153.0"
+VERSION = "2.155.0"
 
 import base64
 import os