offline-intelligence 0.1.3__py3-none-any.whl

offline_intelligence/__init__.py

@@ -0,0 +1,278 @@
+"""
+offline_intelligence - Python bindings for Offline Intelligence Library
+Version: 0.1.3
+
+Pure-Python HTTP client that communicates with the Offline Intelligence
+server (default port 9999). Supports streaming via Server-Sent Events.
+"""
+
+from __future__ import annotations
+
+import os
+import json
+from typing import Any, Dict, Generator, Iterator, Optional
+
+try:
+    import requests
+    from requests import Response
+except ImportError as e:
+    raise ImportError(
+        "The 'requests' package is required. Install it with: pip install requests"
+    ) from e
+
+__version__ = "0.1.3"
+__all__ = ["Config", "OfflineIntelligence", "OfflineIntelligenceException"]
+
+
+# ---------------------------------------------------------------------------
+# Exception
+# ---------------------------------------------------------------------------
+
+class OfflineIntelligenceException(Exception):
+    """Raised when the Offline Intelligence server returns an error."""
+    def __init__(self, message: str, status_code: Optional[int] = None):
+        super().__init__(message)
+        self.status_code = status_code
+
+
+# ---------------------------------------------------------------------------
+# Config
+# ---------------------------------------------------------------------------
+
+class Config:
+    """Configuration for the Offline Intelligence HTTP client."""
+
+    def __init__(self):
+        self.model_path: str = "default.gguf"
+        self.llama_bin: str = "llama-server"
+        self.llama_host: str = "127.0.0.1"
+        self.llama_port: int = 8081
+        self.backend_url: str = "http://127.0.0.1:8081"
+        self.openrouter_api_key: str = ""
+        self.ctx_size: int = 8192
+        self.batch_size: int = 256
+        self.threads: int = 6
+        self.gpu_layers: int = 20
+        self.health_timeout_seconds: int = 60
+        self.hot_swap_grace_seconds: int = 25
+        self.max_concurrent_streams: int = 4
+        self.prometheus_port: int = 9000
+        self.api_host: str = "127.0.0.1"
+        self.api_port: int = 9999
+        self.requests_per_second: int = 24
+        self.generate_timeout_seconds: int = 300
+        self.stream_timeout_seconds: int = 600
+        self.health_check_timeout_seconds: int = 90
+        self.queue_size: int = 100
+        self.queue_timeout_seconds: int = 30
+
+    @classmethod
+    def from_env(cls) -> "Config":
+        """Create a Config populated from environment variables."""
+        cfg = cls()
+        if v := os.getenv("MODEL_PATH"):
+            cfg.model_path = v
+        if v := os.getenv("LLAMA_BIN"):
+            cfg.llama_bin = v
+        if v := os.getenv("LLAMA_HOST"):
+            cfg.llama_host = v
+        if v := os.getenv("LLAMA_PORT"):
+            cfg.llama_port = int(v)
+        if v := os.getenv("BACKEND_URL"):
+            cfg.backend_url = v
+        if v := os.getenv("OPENROUTER_API_KEY"):
+            cfg.openrouter_api_key = v
+        if v := os.getenv("API_HOST"):
+            cfg.api_host = v
+        if v := os.getenv("API_PORT"):
+            cfg.api_port = int(v)
+        if v := os.getenv("CTX_SIZE"):
+            cfg.ctx_size = int(v)
+        if v := os.getenv("GPU_LAYERS"):
+            cfg.gpu_layers = int(v)
+        return cfg
+
+    def __repr__(self) -> str:
+        return (
+            f"Config(api_host={self.api_host!r}, api_port={self.api_port}, "
+            f"model_path={self.model_path!r})"
+        )
+
+
+# ---------------------------------------------------------------------------
+# Main Client
+# ---------------------------------------------------------------------------
+
+class OfflineIntelligence:
+    """
+    HTTP client for the Offline Intelligence server.
+
+    Usage::
+
+        from offline_intelligence import Config, OfflineIntelligence
+
+        cfg = Config.from_env()
+        client = OfflineIntelligence(cfg)
+
+        print(client.health_check())
+        print(client.generate("Hello, world!"))
+    """
+
+    def __init__(self, config: Optional[Config] = None):
+        self.config = config or Config()
+        self.base_url = f"http://{self.config.api_host}:{self.config.api_port}"
+        self._session = requests.Session()
+        self._session.headers.update({"Content-Type": "application/json"})
+
+    def _get(self, path: str, timeout: Optional[int] = None) -> Any:
+        url = f"{self.base_url}{path}"
+        try:
+            resp = self._session.get(url, timeout=timeout or self.config.health_timeout_seconds)
+            resp.raise_for_status()
+            return resp.json()
+        except requests.HTTPError as e:
+            raise OfflineIntelligenceException(str(e), e.response.status_code if e.response else None)
+        except requests.RequestException as e:
+            raise OfflineIntelligenceException(str(e))
+
+    def _post(self, path: str, body: Optional[Dict] = None, timeout: Optional[int] = None) -> Any:
+        url = f"{self.base_url}{path}"
+        try:
+            resp = self._session.post(url, json=body, timeout=timeout or self.config.generate_timeout_seconds)
+            resp.raise_for_status()
+            return resp.json()
+        except requests.HTTPError as e:
+            raise OfflineIntelligenceException(str(e), e.response.status_code if e.response else None)
+        except requests.RequestException as e:
+            raise OfflineIntelligenceException(str(e))
+
+    def _delete(self, path: str, timeout: Optional[int] = None) -> Any:
+        url = f"{self.base_url}{path}"
+        try:
+            resp = self._session.delete(url, timeout=timeout or self.config.health_timeout_seconds)
+            resp.raise_for_status()
+            return resp.json()
+        except requests.HTTPError as e:
+            raise OfflineIntelligenceException(str(e), e.response.status_code if e.response else None)
+        except requests.RequestException as e:
+            raise OfflineIntelligenceException(str(e))
+
+    # ── Health & Status ────────────────────────────────────────────────────
+
+    def health_check(self) -> Dict:
+        """GET /healthz — returns server health status."""
+        return self._get("/healthz")
+
+    def get_status(self) -> Dict:
+        """GET /admin/status — returns engine/model status."""
+        return self._get("/admin/status")
+
+    # ── Model Management ───────────────────────────────────────────────────
+
+    def load_model(self, model_path: str) -> Dict:
+        """POST /admin/load — load a model by path."""
+        return self._post("/admin/load", {"model_path": model_path})
+
+    def stop_model(self) -> Dict:
+        """POST /admin/stop — stop the running model."""
+        return self._post("/admin/stop")
+
+    # ── Generation ─────────────────────────────────────────────────────────
+
+    def generate(self, prompt: str, **options: Any) -> Dict:
+        """POST /generate — generate a response (non-streaming)."""
+        body = {"prompt": prompt, **options}
+        return self._post("/generate", body, timeout=self.config.generate_timeout_seconds)
+
+    def generate_stream(self, prompt: str, **options: Any) -> Generator[str, None, None]:
+        """
+        POST /generate/stream — stream a response via Server-Sent Events.
+
+        Yields each text chunk as a string.
+
+        Example::
+
+            for chunk in client.generate_stream("Tell me a story"):
+                print(chunk, end="", flush=True)
+        """
+        url = f"{self.base_url}/generate/stream"
+        body = {"prompt": prompt, **options}
+        try:
+            with self._session.post(
+                url,
+                json=body,
+                stream=True,
+                timeout=self.config.stream_timeout_seconds,
+            ) as resp:
+                resp.raise_for_status()
+                for line in resp.iter_lines(decode_unicode=True):
+                    if not line:
+                        continue
+                    if line.startswith("data: "):
+                        data = line[6:]
+                        if data.strip() == "[DONE]":
+                            return
+                        try:
+                            payload = json.loads(data)
+                            # OpenAI-compatible format
+                            if "choices" in payload:
+                                delta = payload["choices"][0].get("delta", {})
+                                content = delta.get("content", "")
+                                if content:
+                                    yield content
+                            elif "text" in payload:
+                                yield payload["text"]
+                        except json.JSONDecodeError:
+                            yield data
+        except requests.HTTPError as e:
+            raise OfflineIntelligenceException(str(e), e.response.status_code if e.response else None)
+        except requests.RequestException as e:
+            raise OfflineIntelligenceException(str(e))
+
+    # ── Conversations ──────────────────────────────────────────────────────
+
+    def get_conversations(self) -> Dict:
+        """GET /conversations — list all conversations."""
+        return self._get("/conversations")
+
+    def get_conversation(self, conversation_id: str) -> Dict:
+        """GET /conversations/{id} — get a single conversation."""
+        return self._get(f"/conversations/{conversation_id}")
+
+    def delete_conversation(self, conversation_id: str) -> Dict:
+        """DELETE /conversations/{id} — delete a conversation."""
+        return self._delete(f"/conversations/{conversation_id}")
+
+    def get_conversation_title(self, conversation_id: str) -> Dict:
+        """GET /conversations/{id}/title — get conversation title."""
+        return self._get(f"/conversations/{conversation_id}/title")
+
+    def generate_title(self, session_id: str, first_message: str) -> Dict:
+        """POST /generate/title — generate a title for a conversation."""
+        return self._post("/generate/title", {
+            "session_id": session_id,
+            "first_message": first_message,
+        })
+
+    # ── Memory ─────────────────────────────────────────────────────────────
+
+    def get_memory_stats(self, session_id: str) -> Dict:
+        """GET /memory/stats/{session_id} — get memory statistics."""
+        return self._get(f"/memory/stats/{session_id}")
+
+    def optimize_memory(self) -> Dict:
+        """POST /memory/optimize — trigger memory optimization."""
+        return self._post("/memory/optimize")
+
+    def cleanup_memory(self) -> Dict:
+        """POST /memory/cleanup — clean up stale memory entries."""
+        return self._post("/memory/cleanup")
+
+    # ── Version ────────────────────────────────────────────────────────────
+
+    @staticmethod
+    def version() -> str:
+        return __version__
+
+    def __repr__(self) -> str:
+        return f"OfflineIntelligence(base_url={self.base_url!r})"

offline_intelligence-0.1.3.dist-info/METADATA

@@ -0,0 +1,83 @@
+Metadata-Version: 2.4
+Name: offline-intelligence
+Version: 0.1.3
+Summary: Python bindings for Offline Intelligence Library
+Home-page: https://github.com/offline-intelligence/offline-intelligence
+Author: Offline Intelligence Team
+Author-email: team@offlineintelligence.com
+Project-URL: Bug Tracker, https://github.com/offline-intelligence/offline-intelligence/issues
+Project-URL: Source, https://github.com/offline-intelligence/offline-intelligence
+Keywords: llm,ai,offline,intelligence,local-ai,http-client
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Requires-Dist: requests>=2.28.0
+Provides-Extra: streaming
+Requires-Dist: sseclient-ng>=1.0.0; extra == "streaming"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: project-url
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
+
+# Offline Intelligence Python Bindings
+
+Python bindings for the Offline Intelligence Library - High-performance LLM inference engine with memory management capabilities.
+
+## Installation
+
+```bash
+pip install offline-intelligence
+```
+
+## Quick Start
+
+```python
+from offline_intelligence import Config, run_server
+
+# Configure the engine
+config = Config.from_env()
+
+# Start the server
+run_server(config)
+```
+
+## Features
+
+- **Core LLM Integration**: Direct access to LLM engine functionality
+- **Memory Management**: Base memory operations and database access
+- **Configuration**: Flexible configuration system
+- **Metrics**: Performance monitoring and telemetry
+- **Proxy Interface**: Stream generation and API proxy functionality
+
+## Architecture
+
+This package provides bindings to the core open-source components (80%) of the Offline Intelligence system. Proprietary extensions are available separately.
+
+## Platform Support
+
+- Windows (x64)
+- macOS (Intel/Apple Silicon)
+- Linux (x64, ARM64)
+
+## License
+
+Apache 2.0

offline_intelligence-0.1.3.dist-info/RECORD

@@ -0,0 +1,5 @@
+offline_intelligence/__init__.py,sha256=VuNZsY84gVhJVQhmmhZXwlzgx5T1zaEPo9Rrd83b4dE,11747
+offline_intelligence-0.1.3.dist-info/METADATA,sha256=a7DOxp64KA9AayFTfrU_VeTJEXAPyjsuNatyUxm2-sk,2688
+offline_intelligence-0.1.3.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+offline_intelligence-0.1.3.dist-info/top_level.txt,sha256=qQOPMXLBQPADgwsI0Ta0hYwYNqTnD7kFSTqaW5tIzDs,21
+offline_intelligence-0.1.3.dist-info/RECORD,,

offline_intelligence-0.1.3.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (82.0.1)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

offline_intelligence-0.1.3.dist-info/top_level.txt

@@ -0,0 +1 @@
+offline_intelligence