Extension configuration¶
The extension currently provides the following configuration options:
- provider¶
- Type
string (FQCN)
- Default
EliasHaeussler\Typo3Solver\ProblemSolving\Solution\Provider\OpenAISolutionProvider
Default solution provider to be used for problem solving.
Note
Custom solutions providers must implement
EliasHaeussler\Typo3Solver\ProblemSolving\Solution\Provider\SolutionProvider
.
- prompt¶
- Type
string (FQCN)
- Default
EliasHaeussler\Typo3Solver\ProblemSolving\Solution\Prompt\DefaultPrompt
Default prompt to be used for problem solving.
Note
Custom prompts must implement
EliasHaeussler\Typo3Solver\ProblemSolving\Solution\Prompt\Prompt
.
- ignoredCodes¶
- Type
string (comma-separated list)
Comma-separated list of exception codes to ignore during problem solving. Once a raised exception matches with the configured exception codes, the solution provider won't provide a solution and is simply ignored.
Note
This setting only applies to the default
EliasHaeussler\Typo3Solver\ProblemSolving\Solution\Provider\OpenAISolutionProvider
.
- api.key¶
- Type
string
API key for OpenAI requests.
Attention
It is essential to configure the API key. Otherwise, requests to OpenAI won't be possible.
- attributes.model¶
- Type
string
- Default
gpt-3.5-turbo-0301
OpenAI model to use (see List available models to show a list of available models).
- attributes.maxTokens¶
- Type
integer
- Default
300
Maximum number of tokens to use per request to OpenAI.
- attributes.temperature¶
- Type
float
- Default
0.5
Temperature to use for OpenAI completion requests (must be a value between
0
and1
).
- attributes.numberOfCompletions¶
- Type
integer
- Default
1
Number of completions to generate for each problem.
- cache.lifetime¶
- Type
integer
- Default
86400
(= 1 day)
Lifetime in seconds of the solutions cache.
Tip
Use
0
to disable caching.