.png)
11 hours ago
2
Version: 0.3.0
A C library for interacting with OpenAI-compatible, Google Gemini, and Anthropic Claude LLM APIs, with a focus on creating delightful chaos. Supports text and multimodal inputs, and both regular and streaming responses.
This library is named "Disaster Party" and uses the dp_ prefix for its C symbols. The library file is libdisasterparty.so.
This project uses GNU Autotools.
- Run ./autogen.sh to generate the configure script and Makefiles.
- Run ./configure to check for dependencies and prepare the build.
- You might need to specify a prefix: ./configure --prefix=/usr/local
- Run make to compile the library and tests.
- (Optional) Run make check to run the test suite. See the "Running Tests" section below for details on required environment variables.
- (Optional) Run sudo make install to install the library and header files.
- libcurl (>= 7.20.0)
- libcjson (>= 1.7.10)
Include disasterparty.h in your C code and link against -ldisasterparty. Ensure your build system also links against libcurl and libcjson (pkg-config can help with this).
Example compilation:
See the files in the tests/ directory for usage examples.
The test suite can be run with make check. The tests make live API calls and require certain environment variables to be set. Tests for a specific provider or feature will be skipped if their required variables are not found.
To run the full test suite, the following environment variables are needed:
- OPENAI_API_KEY: Your API key for OpenAI or any OpenAI-compatible service.
- GEMINI_API_KEY: Your API key for Google's Gemini API.
- ANTHROPIC_API_KEY: Your API key for Anthropic's Claude API.
- GEMINI_TEST_IMAGE_PATH: A valid path to a JPEG or PNG image for Gemini multimodal tests.
- ANTHROPIC_TEST_IMAGE_PATH: A valid path to a JPEG or PNG image for Anthropic multimodal tests.
- TEST_IMAGE_PATH_1: Path to the first image for the generic inline context test.
- TEST_IMAGE_PATH_2: Path to the second image for the generic inline context test.
- OPENAI_API_BASE_URL: To test against a custom OpenAI-compatible endpoint instead of the default https://api.openai.com/v1.
The test suite currently includes the following 21 test programs:
- test_openai_text_dp
- test_openai_multimodal_dp
- test_openai_streaming_dp
- test_openai_list_models_dp
- test_openai_streaming_multimodal_dp
- test_gemini_text_dp
- test_gemini_multimodal_dp
- test_gemini_streaming_dp
- test_gemini_list_models_dp
- test_gemini_streaming_multimodal_dp
- test_anthropic_text_dp
- test_anthropic_multimodal_dp
- test_anthropic_streaming_dp
- test_anthropic_streaming_detailed_dp
- test_anthropic_list_models_dp
- test_anthropic_streaming_multimodal_dp
- test_serialization_dp
- test_inline_multimodal_dp
- test_error_handling_dp
- test_parameters_dp
- test_anthropic_streaming_multimodal_detailed_dp
