v1.0.9

v1.0.9

• Added variable maximum tokens depending on openai models
• Micro optimizations (commit 649841e)

v1.0.7

Fixes

• Fixed error when no configuration file specified. Pointed out in issue #2 (commit 781b46b)
• Removed redundant method OpenAIChatCompletion (commit 76196aa)
• Removed unused static method (commit fd87615)
• Handle missing .env file (commit 45976ac)
• Check whether or not destination directory is contained in the source directory (commit 58b580c and commit b722e4e)

v1.0.6

Support for Chat Completion Models

In the closed beta for gpt-4, all 8k or 32k models are chat completion models.
This library now supports chat completion models.

What gpt-4 models bring to the table

Previous models were not able to generate other documentation syntax than JSDOC, the new gpt-4 models can now easily generate typedoc syntax.
Feel free to test out other languages and documentation frameworks (doxygen with cpp, or even rustdoc with rust)

New Pricing System

New models have different prices for prompt tokens and response tokens.
This is now handled. Price is inferred from the model's name, and information about each models is stored in ./src/gpt.ts in the Models lookup table. Feel free to add other model's to this lookup table.

Installation

npm install gptdoc@1.0.6 --save-dev