davinci-resolve-mcp 2.98.2 → 2.98.3

package/CHANGELOG.md

@@ -2,6 +2,65 @@
 
 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 v2.98.3
+
+**`fusion_comp` could never delete a Fusion keyframe.** Reported in
+[#155](https://github.com/samuelgursky/davinci-resolve-mcp/issues/155) by
+@Andrei-59, with the root cause already identified: the handler called a method
+that does not exist. The diagnosis was correct, and the suggested replacement is
+confirmed here against a live build.
+
+### Fixed
+
+- **`delete_keyframe` called `RemoveKeyFrame()` on a Fusion Input.** No such
+  method exists there. Keyframes do not live on the Input — they live on the
+  spline modifier connected to it, which is what `add_keyframe` attaches via
+  `AddModifier(input_name, "BezierSpline")`. The action now reaches that spline
+  through `inp.GetConnectedOutput().GetTool()` and calls `DeleteKeyFrames(time)`
+  on it. Introduced with the tool in v2.1.0 and broken for every input, every
+  frame, and every tool since; there is no version of the server in which it
+  worked.
+
+- **The failure surfaced as `'NoneType' object is not callable`.** The
+  fusionscript bridge resolves an unknown attribute to `None` instead of raising
+  `AttributeError`, so the bad lookup succeeded silently and only died at the
+  callsite — an error naming neither the method nor the object. Every branch of
+  the action now returns the normal error envelope: `FUSION_INPUT_NOT_ANIMATED`
+  when the input has no modifier, `FUSION_KEYFRAME_NOT_FOUND` when nothing is
+  keyed at that frame (the frame list is included in `state`),
+  `FUSION_DELETE_KEYFRAMES_UNSUPPORTED` when the modifier has no removal method,
+  and `INVALID_FRAME` for a non-numeric `time`. The existing `_has_method` guard
+  — which exists precisely for this silent-`None` class — is applied before the
+  call rather than after it.
+
+- **Success is verified by readback, not by the return value.** Live testing
+  showed `DeleteKeyFrames()` returns `None` whether or not it removed anything,
+  so trusting the return would have reported failure on every successful
+  delete — and trusting the absence of an exception would have reported success
+  on every silent no-op. The handler re-reads the keyframe list and returns
+  `FUSION_DELETE_KEYFRAME_NOOP` if the frame is still there. On success it
+  returns `{success, time, remaining_keyframes}`.
+
+### Testing
+
+- `tests/test_fusion_comp_targeting.py` gains eleven `delete_keyframe` cases,
+  including a regression test that models the bridge's silent-`None` attribute
+  lookup and asserts the handler never reaches for `RemoveKeyFrame` on the
+  Input. The action previously had no test coverage at all.
+- `tests/live_fusion_delete_keyframe_validation.py` is a new self-contained live
+  harness: it creates a scratch project, inserts a Fusion composition clip (no
+  media needed), and reports which keyframe methods the Input and the spline
+  actually expose before asserting the delete.
+
+### Verified live
+
+Validated on **DaVinci Resolve Studio 19.1.3.7** (macOS). `RemoveKeyFrame`
+confirmed absent on the Input and resolving to `None`; `DeleteKeyFrames`
+confirmed present on the `BezierSpline` and confirmed to remove the key. The
+reporter's exact reproduction — `add_keyframe` then `delete_keyframe` on the
+same tool/input/frame — was run end-to-end through the patched handler and
+succeeds. Not re-verified on Studio 21.0.4.5, the reporter's build.
+
 ## What's New in v2.98.2
 
 **A tool installed next to the server was invisible to it.** Reported in

package/README.md

@@ -2,7 +2,7 @@
 
 English | [简体中文](README.zh-CN.md)
 
-[![Version](https://img.shields.io/badge/version-2.98.2-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
+[![Version](https://img.shields.io/badge/version-2.98.3-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-35%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-2.98.2-blue.svg)](https://github.com/samuelgursky/davinci-resolve-mcp/releases)
+[![Version](https://img.shields.io/badge/version-2.98.3-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-35%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 v2.98.2 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 v2.98.3 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 = "2.98.2"
+VERSION = "2.98.3"
 # 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": "2.98.2",
+  "version": "2.98.3",
   "description": "NPM bootstrapper for the DaVinci Resolve MCP Server.",
   "license": "MIT",
   "author": "Samuel Gursky <samgursky@gmail.com>",

package/src/granular/common.py

@@ -87,7 +87,7 @@ if not logging.getLogger().handlers:
         handlers=[logging.StreamHandler()],
     )
 
-VERSION = "2.98.2"
+VERSION = "2.98.3"
 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 = "2.98.2"
+VERSION = "2.98.3"
 
 import base64
 import os
@@ -26582,6 +26582,121 @@ def _fusion_set_text_plus(comp, p: Dict[str, Any]) -> Dict[str, Any]:
         comp.Unlock()
 
 
+def _fusion_keyframe_frames(inp) -> List[float]:
+    """Frame positions currently keyed on `inp`, as a sorted list.
+
+    Fusion's `GetKeyFrames()` returns {1-based index: frame_position}; the
+    frames are the VALUES, not the keys. Frames come back as floats.
+    """
+    try:
+        kfs = inp.GetKeyFrames()
+    except Exception:
+        return []
+    if not kfs:
+        return []
+    return sorted(float(frame) for frame in kfs.values())
+
+
+def _fusion_input_spline(inp):
+    """The modifier/spline tool driving `inp`, or None when it is not animated.
+
+    Keyframes do not live on the Input object -- they live on the spline
+    connected to it, which is what `add_keyframe` attaches via AddModifier.
+    """
+    try:
+        connected = inp.GetConnectedOutput()
+    except Exception:
+        return None
+    if connected is None:
+        return None
+    if not _has_method(connected, "GetTool"):
+        return None
+    try:
+        return connected.GetTool()
+    except Exception:
+        return None
+
+
+def _fusion_delete_keyframe(tool, p: Dict[str, Any]) -> Dict[str, Any]:
+    """Remove one keyframe from an animated Fusion input. (issue #155)
+
+    The original implementation called `inp.RemoveKeyFrame(time)`. No such
+    method exists on a Fusion Input, and the fusionscript bridge resolves an
+    unknown attribute to None rather than raising AttributeError -- so the
+    lookup succeeded silently and every call died at the callsite as an opaque
+    `'NoneType' object is not callable`. The action had never worked.
+
+    Deletion happens on the spline, reached the same way `add_keyframe`
+    created it, and every step that can be absent is checked before it is
+    called. The result is verified by reading the keyframe list back, because
+    a Fusion call returning without error is not proof it did anything.
+    """
+    tool_name = p["tool_name"]
+    input_name = p["input_name"]
+    inp = tool[input_name]
+    if not inp:
+        return _err(
+            f"Input '{input_name}' not found on tool '{tool_name}'",
+            code="FUSION_INPUT_NOT_FOUND", category="invalid_input",
+        )
+
+    try:
+        time = float(p["time"])
+    except (TypeError, ValueError):
+        return _err(
+            f"time must be a frame number, got {p['time']!r}",
+            code="INVALID_FRAME", category="invalid_input",
+        )
+
+    spline = _fusion_input_spline(inp)
+    if spline is None:
+        return _err(
+            f"Input '{input_name}' on tool '{tool_name}' is not animated, so it "
+            "has no keyframe to delete",
+            code="FUSION_INPUT_NOT_ANIMATED", category="precondition",
+            remediation="Use add_keyframe first; it attaches the spline that holds keyframes.",
+            state={"tool_name": tool_name, "input_name": input_name},
+        )
+
+    if not _has_method(spline, "DeleteKeyFrames"):
+        return _err(
+            f"The modifier on '{tool_name}.{input_name}' has no DeleteKeyFrames method",
+            code="FUSION_DELETE_KEYFRAMES_UNSUPPORTED", category="unsupported",
+            reason="Only spline modifiers (e.g. BezierSpline) support keyframe removal.",
+            state={"tool_name": tool_name, "input_name": input_name},
+        )
+
+    before = _fusion_keyframe_frames(inp)
+    if not any(abs(frame - time) < 1e-6 for frame in before):
+        return _err(
+            f"No keyframe at frame {time:g} on '{tool_name}.{input_name}'",
+            code="FUSION_KEYFRAME_NOT_FOUND", category="precondition",
+            state={"tool_name": tool_name, "input_name": input_name,
+                   "time": time, "keyframes": before},
+        )
+
+    try:
+        spline.DeleteKeyFrames(time)
+    except Exception as exc:
+        return _err(
+            f"DeleteKeyFrames({time:g}) raised: {exc}",
+            code="FUSION_DELETE_KEYFRAME_FAILED", category="resolve_api_failed",
+            state={"tool_name": tool_name, "input_name": input_name, "time": time},
+        )
+
+    after = _fusion_keyframe_frames(inp)
+    if any(abs(frame - time) < 1e-6 for frame in after):
+        return _err(
+            f"DeleteKeyFrames({time:g}) returned without error but the keyframe "
+            f"is still on '{tool_name}.{input_name}'",
+            code="FUSION_DELETE_KEYFRAME_NOOP", category="resolve_api_failed",
+            state={"tool_name": tool_name, "input_name": input_name,
+                   "time": time, "keyframes_before": before, "keyframes_after": after},
+        )
+
+    return _ok(time=time, remaining_keyframes=after)
+
+
 def _fusion_get_text_plus(comp, p: Dict[str, Any]) -> Dict[str, Any]:
     """Read the text of a Fusion Text+ tool / title template. (issue #73)"""
     tool, err = _fusion_find_text_tool(comp, p)
@@ -26627,7 +26742,9 @@ def fusion_comp(action: str, params: Optional[Dict[str, Any]] = None) -> Dict[st
       get_attrs(tool_name) -> {attrs}
       add_keyframe(tool_name, input_name, time, value) -> {success}
       get_keyframes(tool_name, input_name) -> {keyframes}
-      delete_keyframe(tool_name, input_name, time) -> {success}
+      delete_keyframe(tool_name, input_name, time) -> {success, time, remaining_keyframes}
+        Deletes on the spline attached to the input. Structured errors when the
+        input is not animated or has no keyframe at that frame.
       get_comp_info() -> {name, tool_count, attrs}
       get_position(tool_name) -> {tool_name, x, y}  — read a node's FlowView position
       set_position(tool_name, x, y) -> {success, x, y, readback}  — move a node
@@ -26923,11 +27040,7 @@ def fusion_comp(action: str, params: Optional[Dict[str, Any]] = None) -> Dict[st
             return _err(f"Tool '{p['tool_name']}' not found")
         comp.Lock()
         try:
-            inp = tool[p["input_name"]]
-            if not inp:
-                return _err(f"Input '{p['input_name']}' not found on tool '{p['tool_name']}'")
-            inp.RemoveKeyFrame(p["time"])
-            return _ok()
+            return _fusion_delete_keyframe(tool, p)
         finally:
             comp.Unlock()