feat: added youtube tool (#116)

* feat: added youtube tool

* fix: fixed excessive whitespace in patch example prompt

gptme/tools/__init__.py

@@ -15,6 +15,7 @@
 from .shell import tool as shell_tool
 from .subagent import tool as subagent_tool
 from .tmux import tool as tmux_tool
+from .youtube import tool as youtube_tool
 
 logger = logging.getLogger(__name__)
 
@@ -37,6 +38,7 @@
     browser_tool,
     gh_tool,
     chats_tool,
+    youtube_tool,
     # python tool is loaded last to ensure all functions are registered
     get_python_tool,
 ]

gptme/tools/youtube.py

@@ -0,0 +1,42 @@
+import logging
+
+from ..llm import summarize
+from .base import ToolSpec
+
+logger = logging.getLogger(__name__)
+
+try:
+    # noreorder
+    from youtube_transcript_api import YouTubeTranscriptApi  # fmt: skip
+except ImportError:
+    YouTubeTranscriptApi = None
+    logger.warning(
+        "youtube_transcript_api not available. YouTube tool will be disabled."
+    )
+
+
+def get_transcript(video_id: str) -> str:
+    if not YouTubeTranscriptApi:
+        return "Error: youtube_transcript_api is not installed."
+    try:
+        transcript = YouTubeTranscriptApi.get_transcript(video_id)
+        return " ".join([entry["text"] for entry in transcript])
+    except Exception as e:
+        logger.error(f"Error fetching transcript: {e}")
+        return f"Error fetching transcript: {e}"
+
+
+def summarize_transcript(transcript: str) -> str:
+    return summarize(transcript).content
+
+
+tool: ToolSpec = ToolSpec(
+    name="youtube",
+    desc="Fetch and summarize YouTube video transcripts",
+    instructions="""
+    To use this tool, provide a YouTube video ID and specify whether you want the transcript or a summary.
+    """,
+    functions=[get_transcript, summarize_transcript],
+    block_types=["youtube"],
+    available=bool(YouTubeTranscriptApi),
+)

pyproject.toml

@@ -37,6 +37,9 @@ playwright = {version = "^1.37.0", optional=true}
 openai = "^1.0"
 anthropic = "^0.34.0"
 
+# tools
+youtube_transcript_api = {version = "^0.6.1", optional = true}
+
 # datascience essentials
 matplotlib = {version = "*", optional=true}
 pandas = {version = "*", optional=true}