summariser using custom models added

.gitignore

@@ -133,4 +133,6 @@ dmypy.json
 # Pyre type checker
 .pyre/
 .username
-.idea
+.idea
+*.json
+output/

app/commands/__init__.py

@@ -1,3 +1,5 @@
 from .curator import commands as curator
 from .server import server
 from .media import commands as media
+from .summary import summary
+__all__ = ["media", "curator", "server", "summary"]

app/commands/summary.py

@@ -0,0 +1,103 @@
+import os
+import click
+from ..services.summary import correct_transcript, summarize_transcript, configure_api
+from ..config import settings
+
+def do_correct(input, output, provider, model, api_key):
+    """Implementation logic for transcript correction"""
+    if not api_key:
+        api_key = settings.config.get(f"{provider}_api_key", None)
+
+    configure_api(provider=provider, api_key=api_key)
+
+    with open(input, 'r', encoding='utf-8') as f:
+        original_text = f.read()
+
+    corrected_text = correct_transcript(original_text, provider=provider, model_name=model)
+
+    if corrected_text:
+        if not output:
+            base_name = os.path.splitext(os.path.basename(input))[0]
+            input_dir = os.path.dirname(input) or "."
+            output = os.path.join(input_dir, f"{base_name}_corrected.txt")
+
+        os.makedirs(os.path.dirname(output), exist_ok=True)
+        with open(output, 'w', encoding='utf-8') as f:
+            f.write(corrected_text)
+        click.echo(f"Corrected transcript saved to: {output}")
+
+def do_summarize(input, output, provider, model, api_key, correct_first):
+    """Implementation logic for transcript summarization"""
+    if not api_key:
+        api_key = settings.config.get(f"{provider}_api_key", None)
+
+    configure_api(provider=provider, api_key=api_key)
+
+    with open(input, 'r', encoding='utf-8') as f:
+        original_text = f.read()
+
+    text_to_summarize = original_text
+
+    if correct_first:
+        click.echo("Correcting transcript before summarization...")
+        text_to_summarize = correct_transcript(original_text, provider=provider, model_name=model)
+
+    summary_text = summarize_transcript(text_to_summarize, provider=provider, model_name=model)
+
+    if summary_text:
+        if not output:
+            base_name = os.path.splitext(os.path.basename(input))[0]
+            input_dir = os.path.dirname(input) or "."
+            output = os.path.join(input_dir, f"{base_name}_summary.txt")
+
+        os.makedirs(os.path.dirname(output), exist_ok=True)
+        with open(output, 'w', encoding='utf-8') as f:
+            f.write(summary_text)
+        click.echo(f"Summary saved to: {output}")
+
+@click.group()
+def summary():
+    """Transcript correction and summarization commands."""
+    pass
+
+@summary.command()
+@click.option("--input", "-i", required=True, help="Path to input transcript file.")
+@click.option("--output", "-o", help="Path to save corrected transcript.")
+@click.option("--provider", default="gemini", type=click.Choice(["gemini", "openai", "claude"]), help="AI provider to use.")
+@click.option("--model", help="Model name to use with the selected provider.")
+@click.option("--api-key", help="Custom API key for the selected provider.")
+def correct(input, output, provider, model, api_key):
+    """Correct transcription errors in the transcript."""
+    do_correct(input, output, provider, model, api_key)
+
+@summary.command()
+@click.option("--input", "-i", required=True, help="Path to input transcript file.")
+@click.option("--output", "-o", help="Path to save summary.")
+@click.option("--provider", default="gemini", type=click.Choice(["gemini", "openai", "claude"]), help="AI provider to use.")
+@click.option("--model", help="Model name to use with the selected provider.")
+@click.option("--api-key", help="Custom API key for the selected provider.")
+@click.option("--correct-first/--no-correct", default=False, help="Correct transcript before summarizing.")
+def summarize(input, output, provider, model, api_key, correct_first):
+    """Generate a summary of the transcript."""
+    do_summarize(input, output, provider, model, api_key, correct_first)
+
+@click.command(name="correct")
+@click.option("--input", "-i", required=True, help="Path to input transcript file.")
+@click.option("--output", "-o", help="Path to save corrected transcript.")
+@click.option("--provider", default="gemini", type=click.Choice(["gemini", "openai", "claude"]), help="AI provider to use.")
+@click.option("--model", help="Model name to use with the selected provider.")
+@click.option("--api-key", help="Custom API key for the selected provider.")
+def correct_command(input, output, provider, model, api_key):
+    """Correct transcription errors in the transcript."""
+    do_correct(input, output, provider, model, api_key)
+
+@click.command(name="summarize")
+@click.option("--input", "-i", required=True, help="Path to input transcript file.")
+@click.option("--output", "-o", help="Path to save summary.")
+@click.option("--provider", default="gemini", type=click.Choice(["gemini", "openai", "claude"]), help="AI provider to use.")
+@click.option("--model", help="Model name to use with the selected provider.")
+@click.option("--api-key", help="Custom API key for the selected provider.")
+@click.option("--correct-first/--no-correct", default=False, help="Correct transcript before summarizing.")
+def summarize_command(input, output, provider, model, api_key, correct_first):
+    """Generate a summary of the transcript."""
+    do_summarize(input, output, provider, model, api_key, correct_first)

app/config.py

@@ -36,7 +36,14 @@ def get_config_overview(self):
         overview += f"GITHUB_METADATA_REPO_NAME: {self.GITHUB_METADATA_REPO_NAME}\n"
         overview += f"TRANSCRIPTION_SERVER_URL: {self.TRANSCRIPTION_SERVER_URL}\n"
         overview += f"BTC_TRANSCRIPTS_URL: {self.BTC_TRANSCRIPTS_URL}\n"
-
+        overview += f"GEMINI_API_KEY: {'[SET]' if self.GEMINI_API_KEY else '[NOT SET]'}\n"
+        overview += f"OPENAI_API_KEY: {'[SET]' if self.OPENAI_API_KEY else '[NOT SET]'}\n"
+        overview += f"ANTHROPIC_API_KEY: {'[SET]' if self.ANTHROPIC_API_KEY else '[NOT SET]'}\n"
+        overview += f"DEFAULT_SUMMARY_PROVIDER: {self.DEFAULT_SUMMARY_PROVIDER}\n"
+        overview += f"DEFAULT_GEMINI_MODEL: {self.DEFAULT_GEMINI_MODEL}\n"
+        overview += f"DEFAULT_OPENAI_MODEL: {self.DEFAULT_OPENAI_MODEL}\n"
+        overview += f"DEFAULT_CLAUDE_MODEL: {self.DEFAULT_CLAUDE_MODEL}\n"
+
         # Add config.ini settings
         overview += "\nSettings from config.ini:\n"
         for key, value in self.config.items():
@@ -80,7 +87,36 @@ def GITHUB_PRIVATE_KEY(self):
     def GITHUB_INSTALLATION_ID(self):
         return self._get_env_variable('GITHUB_INSTALLATION_ID',
             "To use GitHub App integration, you need to define a 'GITHUB_INSTALLATION_ID' in your .env file")
+    # Add these properties to the Settings class in app/config.py
+
+    @property
+    def GEMINI_API_KEY(self):
+        # First check environment, then config file
+        return os.getenv('GEMINI_API_KEY') or self.config.get('gemini_api_key', '')
+
+    @property
+    def OPENAI_API_KEY(self):
+        return os.getenv('OPENAI_API_KEY') or self.config.get('openai_api_key', '')
+
+    @property
+    def ANTHROPIC_API_KEY(self):
+        return os.getenv('ANTHROPIC_API_KEY') or self.config.get('anthropic_api_key', '')
 
+    @property
+    def DEFAULT_SUMMARY_PROVIDER(self):
+        return os.getenv('DEFAULT_SUMMARY_PROVIDER') or self.config.get('default_summary_provider', 'gemini')
+
+    @property
+    def DEFAULT_GEMINI_MODEL(self):
+        return os.getenv('DEFAULT_GEMINI_MODEL') or self.config.get('default_gemini_model', 'gemma-3-27b-it')
+
+    @property
+    def DEFAULT_OPENAI_MODEL(self):
+        return os.getenv('DEFAULT_OPENAI_MODEL') or self.config.get('default_openai_model', 'gpt-4o')
+
+    @property
+    def DEFAULT_CLAUDE_MODEL(self):
+        return os.getenv('DEFAULT_CLAUDE_MODEL') or self.config.get('default_claude_model', 'claude-3-7-sonnet-20250219')
 # Initialize the Settings class and expose an instance
 settings = Settings()
 

app/services/__init__.py

@@ -1,2 +1,7 @@
 from .whisper import Whisper
-from .deepgram import Deepgram
+from .deepgram import Deepgram
+from .summary import (
+    correct_transcript,
+    summarize_transcript,
+    configure_api
+)