HandyLLM v0.9.3

Release HandyLLM v0.9.3.

Fixed

• Due to the change in httpx==0.28.0 (see its CHANGELOG), the request params cannot be {} as it will override the original query params. This will cause errors like 404 Resource Not Found, when the api-version parameter in Azure endpoint is overridden. Now when params is an empty dictionary, it will be passed as the default None. See httpx PR 3364 and Issue 3433.

HandyLLM v0.9.2

Release HandyLLM v0.9.2.

Fixed

• fix DictProxy YAML dump
• fix tool calls in stream mode: missing last tool call

Changed

• change import behavior:
  • __init__.py:
    • use explicit imports instead of *
    • limit imports from utils.py
    • remove all imports from response.py and types.py (please import them directly)
  • response.py: add missing items to __all__
• add py.typed for better IDE support

HandyLLM v0.9.1

Release HandyLLM v0.9.1.

Added

• 🔧 add lint dependency ruff
  • add scripts/lint.sh for linting
  • add scripts/format.sh for formatting, and format all files
  • add CI test workflow

Fixed

• fix errors reported by ruff (E711, E722, F541)
• fix TypeError on python 3.8 ('type' object is not subscriptable)

HandyLLM v0.9.0

Release HandyLLM v0.9.0.

Added

Check 🌟s for attractive new features!

• hprompt.py:
  • 🌟 image_url in chat hprompt file now supports local path (file://), both absolute and relative
  • 🌟 add fetch(), afetch(), stream() and astream() methods for direct and typed API responses
  • add RunConfig.var_map_file_format for specifying variable map file format, including JSON / YAML; load_var_map() supports format param
• requestor.py:
  • 🌟 add fetch(), afetch(), stream() and astream() methods for typed API responses
  • use generic and add DictRequestor, BinRequestor, ChatRequestor and CompletionsRequestor
• OpenAIClient:
  • 🌟 constructor supports endpoint_manager, endpoints and load_path param; supports loading from YAML file and Mapping obj
  • APIs support endpoints param
  • APIs endpoint param supports Mapping type
• 🌟 added cache_manager.py: CacheManager for general purpose caching to text files
  • add load_method and dump_method params
  • infers format from file suffix when convert handler is not provided
• 🌟 added response.py: DictProxy for quick access to well-defined response dict
• EndpointManager: supports loading from YAML file using endpoints key, or from Iterable obj
• __init__.py import everything from hprompt for convenience
• rename _types.py to types.py and expose all definitions
• prompt_converter.py:
  • add generator sink consume_stream2fd()
• utils.py:
  • add generator filter trans_stream_chat(), generator sink echo_consumer()
• a lot improved type hints
• added tests:
  • load prompt type specification
  • variable map substitution
  • ChatPrompt and CompletionsPrompt's API calls, supports for RunConfig.on_chunk, and addition operations
  • chat hprompt image_url
  • OpenAIClient loading, chat fetch() & stream()
  • endpoint_manager.py
  • cache_manager.py
  • audio speech
  • legacy OpenAIAPI

Changed

• hprompt.py:
  • add 'endpoints' to default record blacklist
  • remove the var_map related configurations from the evaluated prompt, as it is already applied
• EndpointManager:
  • raises ValueError when getting endpoint out of empty

Fixed

• hprompt.py: 'type' object is not subscriptable on python 3.8

Removed

• prompt_converter.py: remove stream_msgs2raw() and astream_msgs2raw() as no longer needed

HandyLLM v0.8.2

Release HandyLLM v0.8.2.

Added

• hprompt: load methods now support cls parameter for prompt type specification
• ChatPrompt and CompletionsPrompt support optional request and meta
• ChatPrompt :
  • supports add dict
  • add add_message(...) method
• CompletionsPrompt:
  • add add_text(...) method
• PromptConverter: yaml.dump uses allow_unicode=True option
• move all type definitions to _types.py
• support for package development:
  • add requirement.txt for development
  • add scripts/test.sh for running tests
  • add test scripts in tests folder

Fixed

• HandyPrompt.eval(...) should not make directories for output paths
• CompletionsPrompt._run_with_client(...): misplaced run_config param
• PromptConverter
  • fix variable replacement for content_array message
  • fix wrong return type of stream_msgs2raw and astream_msgs2raw
• requestor:
  • httpx.Response should use reason_phrase to get error reason
  • acall() fix missing brackets for await
  • _call_raw() and _acall_raw() intercept and raise new exception without original one
  • _acall_raw(): read the response first to prevent httpx.ResponseNotRead before getting error message
• _utils.exception2err_msg(...) should append error message instead of printing
• change io.IOBase to IO[str] for file descriptors (e.g. RunConfig.output_fd)
• fix other type hints

Changed

• move all old files in tests folder to examples folder

HandyLLM v0.8.1

Release HandyLLM v0.8.1.

Fixed

• fix the debug print issue when outputting to a file in stream mode

HandyLLM v0.8.0

Release HandyLLM v0.8.0.

Added

• CLI: output to stderr without buffering
• add RunConfig.output_path_buffering for controlling buffering of output file
• add this changelog

Fixed

• fix _post_check_output(...) not using evaluated run_config (may cause output_path or output_fd to be ignored)

Changed

• rename internal constants to remove leading _ of API_xxx constants

Removed

• remove unused files in deprecated folder

HandyLLM v0.7.6

Release HandyLLM v0.7.6.

Added

• add RunConfig.on_chunk as callback for streamed chunks
• add Azure tts example
• add VM method to transform kwargs to % wrapped variable map dict
• add var_map arg to eval(...), run(...) and arun(...) for convenience

Changed

• merging different versions of var_map from method argument or from another RunConfig, instead of replacing it as a whole
• rename RunConfig.to_dict's retain_fd arg to retain_object

HandyLLM v0.7.5

Release HandyLLM v0.7.5.

Added

• OpenAIClient add audio speech (tts) api support
  • add azure support for audio speech and transcriptions
• add tts test script

Changed

• prioritize RunConfig.output_evaled_prompt_fd over RunConfig.output_evaled_prompt_path
• eval(...)
  • always return a new object
  • gives run_config arg a default value
  • accepts kwargs, same as run(...)
• when dumping, always filter request
• credential file do not overwrite existing request args

Fixed

• non-stream mode prioritize RunConfig.output_fd over RunConfig.output_path

HandyLLM v0.7.4

Release HandyLLM v0.7.4.