openai-http-proxy 3.2.3__tar.gz → 3.3.0__tar.gz

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openai-http-proxy
-Version: 3.2.3
+Version: 3.3.0
 Summary: OpenAI HTTP Proxy is an OpenAI-compatible http proxy server for inferencing various LLMs capable of working with Google, Anthropic, OpenAI APIs, local PyTorch inference, etc.
 License: MIT License
          
@@ -46,7 +46,7 @@ Provides-Extra: all
 Provides-Extra: anthropic
 Provides-Extra: google
 Provides-Extra: test
-Requires-Dist: ai-microcore (>=6.2.1,<7)
+Requires-Dist: ai-microcore (>=6.6.1,<7)
 Requires-Dist: anthropic (>=0.77,<1) ; extra == "all"
 Requires-Dist: anthropic (>=0.77,<1) ; extra == "anthropic"
 Requires-Dist: fastapi (>=0.136.3,<1)
@@ -116,6 +116,8 @@ It works as a drop-in replacement for OpenAI's API, allowing you to switch betwe
 - [Guides & Reference](#-guides--reference)
 - [Known Limitations](#-known-limitations)
 - [Debugging](#-debugging)
+  - [Debugging Mode](#debugging-mode)
+  - [Printing LLM Requests and Responses](#printing-llm-requests-and-responses)
 - [Contributing](#-contributing)
 - [License](#-license)
 
@@ -255,6 +257,7 @@ OpenAI HTTP Proxy is configured through a TOML/YAML/JSON/Python file that specif
 host = "0.0.0.0"  # Interface to bind to
 port = 8000       # Port to listen on
 dev_autoreload = false  # Enable for development
+print_stream = false    # Print LLM requests / responses to stdout, see "Debugging"
 
 # API key validation function (optional)
 api_key_check = "lm_proxy.api_key_check.check_api_key_in_config"
@@ -761,7 +764,9 @@ For more detailed information, check out these articles:
 
 ## 🔍 Debugging<a id="-debugging"></a>
 
-### Overview
+### Debugging Mode<a id="debugging-mode"></a>
+
+#### Overview
 When **debugging mode** is enabled,
 OpenAI HTTP Proxy provides detailed logging information to help diagnose issues:
 - Stack traces for exceptions are shown in the console
@@ -770,7 +775,7 @@ OpenAI HTTP Proxy provides detailed logging information to help diagnose issues:
 > **Warning** ⚠️  
 > Never enable debugging mode in production environments, as it may expose sensitive information to the application logs.
 
-### Enabling Debugging Mode
+#### Enabling Debugging Mode
 To enable debugging, set the `LM_PROXY_DEBUG` environment variable to a truthy value (e.g., "1", "true", "yes").
 > **Tip** 💡  
 > Environment variables can also be defined in a `.env` file.
@@ -782,6 +787,30 @@ Alternatively, you can enable or disable debugging via the command-line argument
 > **Note** ℹ️   
 > CLI arguments override environment variable settings.
 
+### Printing LLM Requests and Responses<a id="printing-llm-requests-and-responses"></a>
+
+The `print_stream` configuration option makes OpenAI HTTP Proxy print every LLM request
+and response to stdout, with the response appearing token by token as it is generated:
+
+```toml
+print_stream = true
+```
+
+```text
+Requesting LLM gpt-5:
+    [User]:
+        What is the capital of France?
+LLM Response:
+    The capital of France is Paris.
+```
+
+This works for both streaming and non-streaming client requests.
+
+> **Warning** ⚠️  
+> Intended for local development and troubleshooting only.
+> Prompts and generated content are written to the console in plain text,
+> and the output of requests processed in parallel is interleaved.
+
 
 ## 🤝 Contributing<a id="-contributing"></a>
 

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/README.md

@@ -50,6 +50,8 @@ It works as a drop-in replacement for OpenAI's API, allowing you to switch betwe
 - [Guides & Reference](#-guides--reference)
 - [Known Limitations](#-known-limitations)
 - [Debugging](#-debugging)
+  - [Debugging Mode](#debugging-mode)
+  - [Printing LLM Requests and Responses](#printing-llm-requests-and-responses)
 - [Contributing](#-contributing)
 - [License](#-license)
 
@@ -189,6 +191,7 @@ OpenAI HTTP Proxy is configured through a TOML/YAML/JSON/Python file that specif
 host = "0.0.0.0"  # Interface to bind to
 port = 8000       # Port to listen on
 dev_autoreload = false  # Enable for development
+print_stream = false    # Print LLM requests / responses to stdout, see "Debugging"
 
 # API key validation function (optional)
 api_key_check = "lm_proxy.api_key_check.check_api_key_in_config"
@@ -695,7 +698,9 @@ For more detailed information, check out these articles:
 
 ## 🔍 Debugging<a id="-debugging"></a>
 
-### Overview
+### Debugging Mode<a id="debugging-mode"></a>
+
+#### Overview
 When **debugging mode** is enabled,
 OpenAI HTTP Proxy provides detailed logging information to help diagnose issues:
 - Stack traces for exceptions are shown in the console
@@ -704,7 +709,7 @@ OpenAI HTTP Proxy provides detailed logging information to help diagnose issues:
 > **Warning** ⚠️  
 > Never enable debugging mode in production environments, as it may expose sensitive information to the application logs.
 
-### Enabling Debugging Mode
+#### Enabling Debugging Mode
 To enable debugging, set the `LM_PROXY_DEBUG` environment variable to a truthy value (e.g., "1", "true", "yes").
 > **Tip** 💡  
 > Environment variables can also be defined in a `.env` file.
@@ -716,6 +721,30 @@ Alternatively, you can enable or disable debugging via the command-line argument
 > **Note** ℹ️   
 > CLI arguments override environment variable settings.
 
+### Printing LLM Requests and Responses<a id="printing-llm-requests-and-responses"></a>
+
+The `print_stream` configuration option makes OpenAI HTTP Proxy print every LLM request
+and response to stdout, with the response appearing token by token as it is generated:
+
+```toml
+print_stream = true
+```
+
+```text
+Requesting LLM gpt-5:
+    [User]:
+        What is the capital of France?
+LLM Response:
+    The capital of France is Paris.
+```
+
+This works for both streaming and non-streaming client requests.
+
+> **Warning** ⚠️  
+> Intended for local development and troubleshooting only.
+> Prompts and generated content are written to the console in plain text,
+> and the output of requests processed in parallel is interleaved.
+
 
 ## 🤝 Contributing<a id="-contributing"></a>
 

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/lm_proxy/bootstrap.py

@@ -94,7 +94,10 @@ class Env:
                     env.connections[conn_name] = mc.env().llm_async_function
             except mc.LLMConfigError as e:
                 raise ValueError(f"Error in configuration for connection '{conn_name}': {e}") from e
-
+        if config.print_stream:
+            mc.use_logging(stream=True)
+            # Proxied prompts are printed in full, without "...(output was truncated)..."
+            mc.logging.LoggingConfig.STRIP_REQUEST_LINES = None
         logging.info("Done initializing %d connections.", len(env.connections))
 
 

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/lm_proxy/config.py

@@ -90,6 +90,10 @@ class Config(BaseModel):
     )
     components: dict[str, Union[str, Callable, dict]] = Field(default_factory=dict)
     before: list[Union[str, Callable, dict]] = Field(default_factory=list)
+    print_stream: bool = Field(
+        default=False,
+        description="Prints LLM requests / responses to stdout"
+    )
 
     def __init__(self, **data):
         super().__init__(**data)

openai_http_proxy-3.3.0/lm_proxy/config_loaders/json.py

@@ -0,0 +1,11 @@
+"""JSON configuration loader."""
+
+import json
+
+
+def load_json_config(config_path: str) -> dict:
+    """Loads configuration from a JSON file."""
+    # utf-8-sig skips the byte order mark that Windows text editors may add,
+    # the JSON parser rejects it
+    with open(config_path, "r", encoding="utf-8-sig") as f:
+        return json.load(f)

openai_http_proxy-3.3.0/lm_proxy/config_loaders/toml.py

@@ -0,0 +1,11 @@
+"""TOML configuration loader."""
+
+import tomllib
+
+
+def load_toml_config(config_path: str) -> dict:
+    """Loads configuration from a TOML file."""
+    # utf-8-sig skips the byte order mark that Windows text editors may add,
+    # tomllib rejects it
+    with open(config_path, "r", encoding="utf-8-sig") as f:
+        return tomllib.loads(f.read())

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/lm_proxy/core.py

@@ -11,6 +11,7 @@ import inspect
 from datetime import datetime
 from typing import Optional
 
+import microcore as mc
 from fastapi import HTTPException
 from openai.types.chat.chat_completion import Choice
 from starlette.requests import Request
@@ -65,6 +66,30 @@ def resolve_connection_and_model(config: Config, external_model: str) -> tuple[s
     )
 
 
+def print_llm_request(ctx: RequestContext):
+    """
+    Prints the LLM request to stdout when the `print_stream` option is enabled.
+
+    LM-Proxy queries the LLM API function directly instead of using microcore.allm(),
+    so microcore request / response handlers are triggered explicitly.
+    """
+    if env.config.print_stream:
+        for handler in mc.env().llm_before_handlers:
+            handler(ctx.request.messages, **ctx.llm_params)
+
+
+def finish_printing_llm_response(ctx: RequestContext):
+    """
+    Completes printing of the LLM response to stdout, in particular terminates
+    the line of the streamed output, when the `print_stream` option is enabled.
+
+    See print_llm_request().
+    """
+    if env.config.print_stream:
+        for handler in mc.env().llm_after_handlers:
+            handler(ctx.response)
+
+
 async def process_stream(
     async_llm_func,
     request: ChatCompletionRequest,
@@ -121,6 +146,7 @@ async def process_stream(
         except Exception as e:
             ctx.error = e
             yield make_chunk(error={"message": str(e), "type": type(e).__name__})
+        finish_printing_llm_response(ctx)
 
     yield make_chunk(finish_reason="error" if ctx.error else "stop")
     yield "data: [DONE]\n\n"
@@ -252,6 +278,7 @@ async def chat_completions(request: ChatCompletionRequest, raw_request: Request)
     async_llm_func = env.connections[connection]
 
     logging.info("Querying LLM... params: %s", ctx.llm_params)
+    print_llm_request(ctx)
     if request.stream:
         return StreamingResponse(
             process_stream(async_llm_func, request, ctx),
@@ -261,9 +288,11 @@ async def chat_completions(request: ChatCompletionRequest, raw_request: Request)
     try:
         out = await async_llm_func(request.messages, **ctx.llm_params)
         ctx.response = out
+        finish_printing_llm_response(ctx)
         logging.info("LLM response: %s", out)
     except Exception as e:
         ctx.error = e
+        finish_printing_llm_response(ctx)
         await log_non_blocking(ctx)
         error_details = f" [{type(e).__name__}]: {e}" if env.debug else ""
         logging.exception(e)

{openai_http_proxy-3.2.3 → openai_http_proxy-3.3.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "openai-http-proxy"
-version = "3.2.3"
+version = "3.3.0"
 description = "OpenAI HTTP Proxy is an OpenAI-compatible http proxy server for inferencing various LLMs capable of working with Google, Anthropic, OpenAI APIs, local PyTorch inference, etc."
 readme = "README.md"
 keywords = ["llm", "large language models", "ai", "gpt", "openai", "proxy", "http", "proxy-server", "llm gateway", "openai", "anthropic", "google genai"]
@@ -35,7 +35,7 @@ documentation = "https://github.com/Nayjest/lm-proxy#readme"
 license = { file = "LICENSE" }
 
 dependencies = [
-    "ai-microcore>=6.2.1,<7",
+    "ai-microcore>=6.6.1,<7",
     "fastapi>=0.136.3,<1",
     "uvicorn>=0.41.0",
     "typer>=0.24.0,<1",

openai_http_proxy-3.2.3/lm_proxy/config_loaders/json.py

@@ -1,9 +0,0 @@
-"""JSON configuration loader."""
-
-import json
-
-
-def load_json_config(config_path: str) -> dict:
-    """Loads configuration from a JSON file."""
-    with open(config_path, "r", encoding="utf-8") as f:
-        return json.load(f)

openai_http_proxy-3.2.3/lm_proxy/config_loaders/toml.py

@@ -1,9 +0,0 @@
-"""TOML configuration loader."""
-
-import tomllib
-
-
-def load_toml_config(config_path: str) -> dict:
-    """Loads configuration from a TOML file."""
-    with open(config_path, "rb") as f:
-        return tomllib.load(f)