npcsh 0.3.27.7__tar.gz → 0.3.29__tar.gz

{npcsh-0.3.27.7/npcsh.egg-info → npcsh-0.3.29}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: npcsh
-Version: 0.3.27.7
+Version: 0.3.29
 Summary: npcsh is a command line tool for integrating LLMs into everyday workflows and for orchestrating teams of NPCs.
 Home-page: https://github.com/cagostino/npcsh
 Author: Christopher Agostino
@@ -32,6 +32,7 @@ Requires-Dist: duckduckgo-search
 Requires-Dist: flask
 Requires-Dist: flask_cors
 Requires-Dist: redis
+Requires-Dist: psycopg2-binary
 Requires-Dist: flask_sse
 Provides-Extra: lite
 Requires-Dist: anthropic; extra == "lite"
@@ -468,6 +469,7 @@ if __name__ == "__main__":
 ### Linux install
 ```bash
 
+# for audio primarily
 sudo apt-get install espeak
 sudo apt-get install portaudio19-dev python3-pyaudio
 sudo apt-get install alsa-base alsa-utils
@@ -475,6 +477,10 @@ sudo apt-get install libcairo2-dev
 sudo apt-get install libgirepository1.0-dev
 sudo apt-get install ffmpeg
 
+# for triggers
+sudo apt install inotify-tools
+
+
 #And if you don't have ollama installed, use this:
 curl -fsSL https://ollama.com/install.sh | sh
 
@@ -482,25 +488,46 @@ ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
-```
-If you'd like to install the abilities to use STT and TTS, additionall install the following
-```
-pip install openai-whisper pyaudio gtts playsound
-```
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 
 
 
 ### Mac install
 ```bash
+#mainly for audio
 brew install portaudio
 brew install ffmpeg
+brew install pygobject3
+
+# for triggers
+brew install ...
+
+
 brew install ollama
 brew services start ollama
-brew install pygobject3
 ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 ```
 ### Windows Install
 
@@ -513,6 +540,16 @@ ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 ```
 As of now, npcsh appears to work well with some of the core functionalities like /ots and /whisper.
 
@@ -1050,13 +1087,30 @@ npc ots -f test_data/catfight.PNG
 ### Plan : Schedule tasks to be run at regular intervals (under construction)
 Use the /plan macro to schedule tasks to be run at regular intervals.
 ```npcsh
-npcsh> /plan run a rag search on the files in the current directory every 5 minutes
+npcsh> /plan run a rag search for 'moonbeam' on the files in the current directory every 5 minutes
+```
+
+```npcsh
+npcsh> /plan record the cpu usage every 5 minutes
 ```
 
+```npcsh
+npcsh> /plan record the apps that are using the most ram every 5 minutes
+```
+
+
+
+
 ```bash
 npc plan -f 30m -t 'task'
 ```
 
+Plan will use platform-specific scheduling tools. In particular, it uses crontab on Linux and launchd on macOS and Schedule Tasks on Windows.
+
+Implementations have been provided for Mac and Windows but only has been tested as of 3/23/2025 on Linux.
+
+
+
 ### Plonk : Computer Control
 Use the /plonk macro to allow the LLM to control your computer.
 ```npcsh
@@ -1342,6 +1396,18 @@ npcsh> /spool model=llama3.3
 npc spool -n npc.npc
 ```
 
+### Trigger
+Use the /trigger macro to execute specific actionss based on certain conditions.
+
+```npcsh
+npcsh> /trigger watch for new PDF downloads in the ~/Downloads directory and move them
+to the ~/Documents/PDFs directory . Ensure that the directory exists or create it if it does not.
+```
+
+On Linux, trigger makes use of inotify-tools to watch for file system events. On macOS, it uses fswatch, and on Windows, it uses Watch-Command.
+
+
+
 
 
 ### Vixynt: Image Generation

{npcsh-0.3.27.7 → npcsh-0.3.29}/README.md

@@ -385,6 +385,7 @@ if __name__ == "__main__":
 ### Linux install
 ```bash
 
+# for audio primarily
 sudo apt-get install espeak
 sudo apt-get install portaudio19-dev python3-pyaudio
 sudo apt-get install alsa-base alsa-utils
@@ -392,6 +393,10 @@ sudo apt-get install libcairo2-dev
 sudo apt-get install libgirepository1.0-dev
 sudo apt-get install ffmpeg
 
+# for triggers
+sudo apt install inotify-tools
+
+
 #And if you don't have ollama installed, use this:
 curl -fsSL https://ollama.com/install.sh | sh
 
@@ -399,25 +404,46 @@ ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
-```
-If you'd like to install the abilities to use STT and TTS, additionall install the following
-```
-pip install openai-whisper pyaudio gtts playsound
-```
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 
 
 
 ### Mac install
 ```bash
+#mainly for audio
 brew install portaudio
 brew install ffmpeg
+brew install pygobject3
+
+# for triggers
+brew install ...
+
+
 brew install ollama
 brew services start ollama
-brew install pygobject3
 ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 ```
 ### Windows Install
 
@@ -430,6 +456,16 @@ ollama pull llama3.2
 ollama pull llava:7b
 ollama pull nomic-embed-text
 pip install npcsh
+# if you want to install with the API libraries
+pip install npcsh[lite]
+# if you want the full local package set up (ollama, diffusers, transformers, cuda etc.)
+pip install npcsh[local]
+# if you want to use tts/stt
+pip install npcsh[whisper]
+
+# if you want everything:
+pip install npcsh[all]
+
 ```
 As of now, npcsh appears to work well with some of the core functionalities like /ots and /whisper.
 
@@ -967,13 +1003,30 @@ npc ots -f test_data/catfight.PNG
 ### Plan : Schedule tasks to be run at regular intervals (under construction)
 Use the /plan macro to schedule tasks to be run at regular intervals.
 ```npcsh
-npcsh> /plan run a rag search on the files in the current directory every 5 minutes
+npcsh> /plan run a rag search for 'moonbeam' on the files in the current directory every 5 minutes
+```
+
+```npcsh
+npcsh> /plan record the cpu usage every 5 minutes
 ```
 
+```npcsh
+npcsh> /plan record the apps that are using the most ram every 5 minutes
+```
+
+
+
+
 ```bash
 npc plan -f 30m -t 'task'
 ```
 
+Plan will use platform-specific scheduling tools. In particular, it uses crontab on Linux and launchd on macOS and Schedule Tasks on Windows.
+
+Implementations have been provided for Mac and Windows but only has been tested as of 3/23/2025 on Linux.
+
+
+
 ### Plonk : Computer Control
 Use the /plonk macro to allow the LLM to control your computer.
 ```npcsh
@@ -1259,6 +1312,18 @@ npcsh> /spool model=llama3.3
 npc spool -n npc.npc
 ```
 
+### Trigger
+Use the /trigger macro to execute specific actionss based on certain conditions.
+
+```npcsh
+npcsh> /trigger watch for new PDF downloads in the ~/Downloads directory and move them
+to the ~/Documents/PDFs directory . Ensure that the directory exists or create it if it does not.
+```
+
+On Linux, trigger makes use of inotify-tools to watch for file system events. On macOS, it uses fswatch, and on Windows, it uses Watch-Command.
+
+
+
 
 
 ### Vixynt: Image Generation

{npcsh-0.3.27.7 → npcsh-0.3.29}/npcsh/llm_funcs.py

@@ -17,7 +17,7 @@ import numpy as np
 
 from google.generativeai import types
 import google.generativeai as genai
-
+from sqlalchemy import create_engine
 
 from .npc_sysenv import (
     get_system_message,
@@ -1554,7 +1554,7 @@ def check_output_sufficient(
 
 def process_data_output(
     llm_response: Dict[str, Any],
-    db_conn: sqlite3.Connection,
+    db_conn,
     request: str,
     tables: str = None,
     history: str = None,
@@ -1572,9 +1572,15 @@ def process_data_output(
         if not query:
             return {"response": "No query provided", "code": 400}
 
+        # Create SQLAlchemy engine based on connection type
+        if "psycopg2" in db_conn.__class__.__module__:
+            engine = create_engine("postgresql://caug:gobears@localhost/npc_test")
+        else:
+            engine = create_engine("sqlite:///test_sqlite.db")
+
         if choice == 1:  # Direct answer query
             try:
-                df = pd.read_sql_query(query, db_conn)
+                df = pd.read_sql_query(query, engine)
                 result = check_output_sufficient(
                     request, df, query, model=model, provider=provider, npc=npc
                 )
@@ -1591,7 +1597,7 @@ def process_data_output(
 
         elif choice == 2:  # Exploratory query
             try:
-                df = pd.read_sql_query(query, db_conn)
+                df = pd.read_sql_query(query, engine)
                 extra_context = f"""
                 Exploratory query results:
                 Query: {query}
@@ -1621,7 +1627,7 @@ def process_data_output(
 
 def get_data_response(
     request: str,
-    db_conn: sqlite3.Connection,
+    db_conn,
     tables: str = None,
     n_try_freq: int = 5,
     extra_context: str = None,
@@ -1634,9 +1640,73 @@ def get_data_response(
     """
     Generate a response to a data request, with retries for failed attempts.
     """
+
+    # Extract schema information based on connection type
+    schema_info = ""
+    if "psycopg2" in db_conn.__class__.__module__:
+        cursor = db_conn.cursor()
+        # Get all tables and their columns
+        cursor.execute(
+            """
+            SELECT
+                t.table_name,
+                array_agg(c.column_name || ' ' || c.data_type) as columns,
+                array_agg(
+                    CASE
+                        WHEN tc.constraint_type = 'FOREIGN KEY'
+                        THEN kcu.column_name || ' REFERENCES ' || ccu.table_name || '.' || ccu.column_name
+                        ELSE NULL
+                    END
+                ) as foreign_keys
+            FROM information_schema.tables t
+            JOIN information_schema.columns c ON t.table_name = c.table_name
+            LEFT JOIN information_schema.table_constraints tc
+                ON t.table_name = tc.table_name
+                AND tc.constraint_type = 'FOREIGN KEY'
+            LEFT JOIN information_schema.key_column_usage kcu
+                ON tc.constraint_name = kcu.constraint_name
+            LEFT JOIN information_schema.constraint_column_usage ccu
+                ON tc.constraint_name = ccu.constraint_name
+            WHERE t.table_schema = 'public'
+            GROUP BY t.table_name;
+        """
+        )
+        for table, columns, fks in cursor.fetchall():
+            schema_info += f"\nTable {table}:\n"
+            schema_info += "Columns:\n"
+            for col in columns:
+                schema_info += f"  - {col}\n"
+            if any(fk for fk in fks if fk is not None):
+                schema_info += "Foreign Keys:\n"
+                for fk in fks:
+                    if fk:
+                        schema_info += f"  - {fk}\n"
+
+    elif "sqlite3" in db_conn.__class__.__module__:
+        cursor = db_conn.cursor()
+        cursor.execute("SELECT name FROM sqlite_master WHERE type='table';")
+        tables = cursor.fetchall()
+        for (table_name,) in tables:
+            schema_info += f"\nTable {table_name}:\n"
+            cursor.execute(f"PRAGMA table_info({table_name});")
+            columns = cursor.fetchall()
+            schema_info += "Columns:\n"
+            for col in columns:
+                schema_info += f"  - {col[1]} {col[2]}\n"
+
+            cursor.execute(f"PRAGMA foreign_key_list({table_name});")
+            foreign_keys = cursor.fetchall()
+            if foreign_keys:
+                schema_info += "Foreign Keys:\n"
+                for fk in foreign_keys:
+                    schema_info += f"  - {fk[3]} REFERENCES {fk[2]}({fk[4]})\n"
+
     prompt = f"""
     User request: {request}
-    Available tables: {tables or 'Not specified'}
+
+    Database Schema:
+    {schema_info}
+
     {extra_context or ''}
     {f'Query history: {history}' if history else ''}
 
@@ -1655,49 +1725,47 @@ def get_data_response(
 
     failures = []
     for attempt in range(max_retries):
-        try:
-            llm_response = get_llm_response(
-                prompt, npc=npc, format="json", model=model, provider=provider
-            )
+        # try:
+        llm_response = get_llm_response(
+            prompt, npc=npc, format="json", model=model, provider=provider
+        )
 
-            # Clean response if it's a string
-            response_data = llm_response.get("response", {})
-            if isinstance(response_data, str):
-                response_data = (
-                    response_data.replace("```json", "").replace("```", "").strip()
-                )
-                try:
-                    response_data = json.loads(response_data)
-                except json.JSONDecodeError:
-                    failures.append("Invalid JSON response")
-                    continue
-
-            result = process_data_output(
-                response_data,
-                db_conn,
-                request,
-                tables=tables,
-                history=failures,
-                npc=npc,
-                model=model,
-                provider=provider,
+        # Clean response if it's a string
+        response_data = llm_response.get("response", {})
+        if isinstance(response_data, str):
+            response_data = (
+                response_data.replace("```json", "").replace("```", "").strip()
             )
+            try:
+                response_data = json.loads(response_data)
+            except json.JSONDecodeError:
+                failures.append("Invalid JSON response")
+                continue
+
+        result = process_data_output(
+            response_data,
+            db_conn,
+            request,
+            tables=tables,
+            history=failures,
+            npc=npc,
+            model=model,
+            provider=provider,
+        )
 
-            if result["code"] == 200:
-                return result
-
-            failures.append(result["response"])
+        if result["code"] == 200:
+            return result
 
-            if attempt == max_retries - 1:
-                return {
-                    "response": f"Failed after {max_retries} attempts. Errors: {'; '.join(failures)}",
-                    "code": 400,
-                }
+        failures.append(result["response"])
 
-        except Exception as e:
-            failures.append(str(e))
+        if attempt == max_retries - 1:
+            return {
+                "response": f"Failed after {max_retries} attempts. Errors: {'; '.join(failures)}",
+                "code": 400,
+            }
 
-    return {"response": "Max retries exceeded", "code": 400}
+    # except Exception as e:
+    #    failures.append(str(e))
 
 
 def enter_reasoning_human_in_the_loop(

{npcsh-0.3.27.7 → npcsh-0.3.29}/npcsh/npc_compiler.py

@@ -788,11 +788,29 @@ class NPC:
         self.model = model
         self.db_conn = db_conn
         if self.db_conn is not None:
-            self.tables = self.db_conn.execute(
-                "SELECT name, sql FROM sqlite_master WHERE type='table';"
-            ).fetchall()
+            # Determine database type
+            if "psycopg2" in self.db_conn.__class__.__module__:
+                # PostgreSQL connection
+                cursor = self.db_conn.cursor()
+                cursor.execute(
+                    """
+                    SELECT table_name, obj_description((quote_ident(table_name))::regclass, 'pg_class')
+                    FROM information_schema.tables
+                    WHERE table_schema='public';
+                """
+                )
+                self.tables = cursor.fetchall()
+                self.db_type = "postgres"
+            elif "sqlite3" in self.db_conn.__class__.__module__:
+                # SQLite connection
+                self.tables = self.db_conn.execute(
+                    "SELECT name, sql FROM sqlite_master WHERE type='table';"
+                ).fetchall()
+                self.db_type = "sqlite"
         else:
             self.tables = None
+            self.db_type = None
+
         self.provider = provider
         self.api_url = api_url
         self.all_tools = all_tools or []
@@ -839,6 +857,45 @@ class NPC:
         else:
             self.parsed_npcs = []
 
+    def execute_query(self, query, params=None):
+        """Execute a query based on database type"""
+        if self.db_type == "postgres":
+            cursor = self.db_conn.cursor()
+            cursor.execute(query, params or ())
+            return cursor.fetchall()
+        else:  # sqlite
+            cursor = self.db_conn.execute(query, params or ())
+            return cursor.fetchall()
+
+    def _determine_db_type(self):
+        """Determine if the connection is PostgreSQL or SQLite"""
+        # Check the connection object's class name
+        conn_type = self.db_conn.__class__.__module__.lower()
+
+        if "psycopg" in conn_type:
+            return "postgres"
+        elif "sqlite" in conn_type:
+            return "sqlite"
+        else:
+            raise ValueError(f"Unsupported database type: {conn_type}")
+
+    def _get_tables(self):
+        """Get table information based on database type"""
+        if self.db_type == "postgres":
+            cursor = self.db_conn.cursor()
+            cursor.execute(
+                """
+                SELECT table_name, obj_description((quote_ident(table_name))::regclass, 'pg_class') as description
+                FROM information_schema.tables
+                WHERE table_schema='public';
+            """
+            )
+            return cursor.fetchall()
+        else:  # sqlite
+            return self.db_conn.execute(
+                "SELECT name, sql FROM sqlite_master WHERE type='table';"
+            ).fetchall()
+
     def get_memory(self):
         return