agentvibes 5.3.0 → 5.5.0

package/.claude/hooks/soprano-gradio-synth.py

@@ -1,139 +1,139 @@
-#!/usr/bin/env python3
-#
-# File: .claude/hooks/soprano-gradio-synth.py
-#
-# AgentVibes - Finally, your AI Agents can Talk Back!
-# Website: https://agentvibes.org
-# Repository: https://github.com/paulpreibisch/AgentVibes
-#
-# Co-created by Paul Preibisch with Claude AI
-# Copyright (c) 2025 Paul Preibisch
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#     http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-#
-# ---
-#
-# Soprano Gradio WebUI synthesizer helper.
-# Calls the Soprano WebUI's Gradio API and saves the result as a WAV file.
-# Uses only Python stdlib (json, sys, urllib) — no extra dependencies.
-#
-# Usage: python3 soprano-gradio-synth.py "text to speak" output.wav [port]
-#
-"""
-Soprano Gradio WebUI synthesizer helper for AgentVibes.
-
-Calls the Soprano WebUI's Gradio API via the Server-Sent Events (SSE) protocol
-and downloads the generated audio as a WAV file.
-
-Flow:
-  1. Submit generation request → get event_id
-  2. Poll SSE stream for audio file URL
-  3. Download WAV file to output path
-
-See: https://github.com/ekwek1/soprano
-"""
-import json
-import sys
-import urllib.request
-import urllib.error
-
-
-def synth(text: str, output_path: str, port: int = 7860) -> None:
-    base = f"http://127.0.0.1:{port}"
-
-    # Step 1: Submit generation request
-    # Args: text, temperature, top_p, repetition_penalty, chunk_size, streaming
-    payload = json.dumps({
-        "data": [text, 0.0, 0.95, 1.2, 1, False]
-    }).encode()
-
-    event_id = submit_request(base, payload)
-
-    # Step 2: Poll SSE stream for audio file URL
-    audio_url = poll_for_result(base, event_id)
-
-    # Step 3: Download the audio file
-    download_file(audio_url, output_path)
-
-
-def submit_request(base: str, payload: bytes) -> str:
-    """Submit generation request to Gradio API, return event_id."""
-    for api_base in ["/gradio_api/call", "/call"]:
-        url = f"{base}{api_base}/generate_speech"
-        req = urllib.request.Request(
-            url,
-            data=payload,
-            headers={"Content-Type": "application/json"},
-        )
-        try:
-            with urllib.request.urlopen(req, timeout=2) as resp:
-                return json.loads(resp.read())["event_id"]
-        except urllib.error.URLError:
-            continue
-
-    raise RuntimeError("Could not reach Soprano WebUI API")
-
-
-def poll_for_result(base: str, event_id: str) -> str:
-    """Poll SSE endpoint until audio file URL is returned."""
-    for api_base in ["/gradio_api/call", "/call"]:
-        url = f"{base}{api_base}/generate_speech/{event_id}"
-        req = urllib.request.Request(url)
-        try:
-            with urllib.request.urlopen(req, timeout=120) as resp:
-                for raw_line in resp:
-                    line = raw_line.decode("utf-8", errors="replace").strip()
-                    if not line.startswith("data: "):
-                        continue
-                    try:
-                        parsed = json.loads(line[6:])
-                    except json.JSONDecodeError:
-                        continue
-                    # Response format: [{"path": "...", "url": "...", ...}, "status string"]
-                    if isinstance(parsed, list) and len(parsed) >= 1:
-                        audio = parsed[0]
-                        if isinstance(audio, dict) and "url" in audio:
-                            return audio["url"]
-            break
-        except urllib.error.URLError:
-            continue
-
-    raise RuntimeError("No audio URL in Soprano response")
-
-
-def download_file(url: str, output_path: str) -> None:
-    """Download audio file from Gradio file server."""
-    req = urllib.request.Request(url)
-    with urllib.request.urlopen(req, timeout=30) as resp:
-        with open(output_path, "wb") as f:
-            while True:
-                chunk = resp.read(8192)
-                if not chunk:
-                    break
-                f.write(chunk)
-
-
-if __name__ == "__main__":
-    if len(sys.argv) < 3:
-        print(f"Usage: {sys.argv[0]} \"text\" output.wav [port]", file=sys.stderr)
-        sys.exit(1)
-
-    text = sys.argv[1]
-    output = sys.argv[2]
-    port = int(sys.argv[3]) if len(sys.argv) > 3 else 7860
-
-    try:
-        synth(text, output, port)
-    except Exception as e:
-        print(f"Error: {e}", file=sys.stderr)
-        sys.exit(1)
+#!/usr/bin/env python3
+#
+# File: .claude/hooks/soprano-gradio-synth.py
+#
+# AgentVibes - Finally, your AI Agents can Talk Back!
+# Website: https://agentvibes.org
+# Repository: https://github.com/paulpreibisch/AgentVibes
+#
+# Co-created by Paul Preibisch with Claude AI
+# Copyright (c) 2025 Paul Preibisch
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+# ---
+#
+# Soprano Gradio WebUI synthesizer helper.
+# Calls the Soprano WebUI's Gradio API and saves the result as a WAV file.
+# Uses only Python stdlib (json, sys, urllib) — no extra dependencies.
+#
+# Usage: python3 soprano-gradio-synth.py "text to speak" output.wav [port]
+#
+"""
+Soprano Gradio WebUI synthesizer helper for AgentVibes.
+
+Calls the Soprano WebUI's Gradio API via the Server-Sent Events (SSE) protocol
+and downloads the generated audio as a WAV file.
+
+Flow:
+  1. Submit generation request → get event_id
+  2. Poll SSE stream for audio file URL
+  3. Download WAV file to output path
+
+See: https://github.com/ekwek1/soprano
+"""
+import json
+import sys
+import urllib.request
+import urllib.error
+
+
+def synth(text: str, output_path: str, port: int = 7860) -> None:
+    base = f"http://127.0.0.1:{port}"
+
+    # Step 1: Submit generation request
+    # Args: text, temperature, top_p, repetition_penalty, chunk_size, streaming
+    payload = json.dumps({
+        "data": [text, 0.0, 0.95, 1.2, 1, False]
+    }).encode()
+
+    event_id = submit_request(base, payload)
+
+    # Step 2: Poll SSE stream for audio file URL
+    audio_url = poll_for_result(base, event_id)
+
+    # Step 3: Download the audio file
+    download_file(audio_url, output_path)
+
+
+def submit_request(base: str, payload: bytes) -> str:
+    """Submit generation request to Gradio API, return event_id."""
+    for api_base in ["/gradio_api/call", "/call"]:
+        url = f"{base}{api_base}/generate_speech"
+        req = urllib.request.Request(
+            url,
+            data=payload,
+            headers={"Content-Type": "application/json"},
+        )
+        try:
+            with urllib.request.urlopen(req, timeout=30) as resp:
+                return json.loads(resp.read())["event_id"]
+        except urllib.error.URLError:
+            continue
+
+    raise RuntimeError("Could not reach Soprano WebUI API")
+
+
+def poll_for_result(base: str, event_id: str) -> str:
+    """Poll SSE endpoint until audio file URL is returned."""
+    for api_base in ["/gradio_api/call", "/call"]:
+        url = f"{base}{api_base}/generate_speech/{event_id}"
+        req = urllib.request.Request(url)
+        try:
+            with urllib.request.urlopen(req, timeout=120) as resp:
+                for raw_line in resp:
+                    line = raw_line.decode("utf-8", errors="replace").strip()
+                    if not line.startswith("data: "):
+                        continue
+                    try:
+                        parsed = json.loads(line[6:])
+                    except json.JSONDecodeError:
+                        continue
+                    # Response format: [{"path": "...", "url": "...", ...}, "status string"]
+                    if isinstance(parsed, list) and len(parsed) >= 1:
+                        audio = parsed[0]
+                        if isinstance(audio, dict) and "url" in audio:
+                            return audio["url"]
+            break
+        except urllib.error.URLError:
+            continue
+
+    raise RuntimeError("No audio URL in Soprano response")
+
+
+def download_file(url: str, output_path: str) -> None:
+    """Download audio file from Gradio file server."""
+    req = urllib.request.Request(url)
+    with urllib.request.urlopen(req, timeout=30) as resp:
+        with open(output_path, "wb") as f:
+            while True:
+                chunk = resp.read(8192)
+                if not chunk:
+                    break
+                f.write(chunk)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) < 3:
+        print(f"Usage: {sys.argv[0]} \"text\" output.wav [port]", file=sys.stderr)
+        sys.exit(1)
+
+    text = sys.argv[1]
+    output = sys.argv[2]
+    port = int(sys.argv[3]) if len(sys.argv) > 3 else 7860
+
+    try:
+        synth(text, output, port)
+    except Exception as e:
+        print(f"Error: {e}", file=sys.stderr)
+        sys.exit(1)

package/.claude/hooks/stop.sh

@@ -0,0 +1,63 @@
+#!/usr/bin/env bash
+#
+# File: .claude/hooks/stop.sh
+#
+# AgentVibes Stop Hook - Wrapper that delegates to .agentvibes/
+# This hook runs in LITE mode to extract and speak "Audio Summary:" markers
+#
+
+# Fix locale warnings
+export LC_ALL=C
+
+# Get script directory
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+PROJECT_ROOT="$(cd "$SCRIPT_DIR/../.." && pwd)"
+
+# Path to mode configuration
+MODE_FILE="$PROJECT_ROOT/.agentvibes/config/mode.txt"
+
+# Check if AgentVibes is installed
+AGENTVIBES_HOOKS="$PROJECT_ROOT/.agentvibes/hooks"
+if [[ ! -d "$AGENTVIBES_HOOKS" ]]; then
+  # AgentVibes not installed, exit silently
+  exit 0
+fi
+
+# Read current mode (default to full if not set)
+CURRENT_MODE="full"
+if [[ -f "$MODE_FILE" ]]; then
+  CURRENT_MODE=$(cat "$MODE_FILE" 2>/dev/null | tr -d '[:space:]')
+fi
+
+# Only run stop hook in LITE mode
+# (Full mode uses tool calls for TTS, not stop hooks)
+if [[ "$CURRENT_MODE" == "lite" ]]; then
+  # Claude Code sends JSON on stdin: {"transcript_path": "..."}
+  HOOK_INPUT=$(cat)
+  TRANSCRIPT_PATH=$(echo "$HOOK_INPUT" | jq -r '.transcript_path // empty' 2>/dev/null || true)
+
+  [[ -z "$TRANSCRIPT_PATH" ]] && exit 0
+  [[ ! -f "$TRANSCRIPT_PATH" ]] && exit 0
+
+  # Extract Audio Summary ONLY from the most recent assistant response.
+  # Key insight: we must NOT search earlier responses — if the latest response
+  # has no Audio Summary: marker, we stay silent (no stale replay).
+  SUMMARY=$(jq --slurp -r '
+    [.[] | try select(.message.role == "assistant")] | last // null
+    | if . == null then "" else
+        [.message.content[]? | select(.type == "text") | .text] | join("\n")
+      end
+    | if test("[Aa]udio [Ss]ummary:") then . else "" end
+  ' "$TRANSCRIPT_PATH" 2>/dev/null \
+    | sed -n 's/.*[Aa]udio [Ss]ummary:[[:space:]]*\(.*\)/\1/p' \
+    | head -1 \
+    | sed 's/^[[:space:]]*//;s/[[:space:]]*$//')
+
+  [[ -z "$SUMMARY" ]] && exit 0
+
+  # Speak via AgentVibes play-tts.sh
+  if [[ -f "$PROJECT_ROOT/.claude/hooks/play-tts.sh" ]]; then
+    AGENTVIBES_MIN_TOKENS=3 AGENTVIBES_SHORT_TOKENS=50 \
+      bash "$PROJECT_ROOT/.claude/hooks/play-tts.sh" "$SUMMARY" >/dev/null 2>&1 &
+  fi
+fi