Skip to content

Generation

Init file.

Generation #

Bases: BaseSynthesizer

Source code in llama-index-core/llama_index/core/response_synthesizers/generation.py
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
class Generation(BaseSynthesizer):
    def __init__(
        self,
        llm: Optional[LLMPredictorType] = None,
        callback_manager: Optional[CallbackManager] = None,
        prompt_helper: Optional[PromptHelper] = None,
        simple_template: Optional[BasePromptTemplate] = None,
        streaming: bool = False,
        # deprecated
        service_context: Optional[ServiceContext] = None,
    ) -> None:
        if service_context is not None:
            prompt_helper = service_context.prompt_helper

        super().__init__(
            llm=llm,
            callback_manager=callback_manager,
            prompt_helper=prompt_helper,
            service_context=service_context,
            streaming=streaming,
        )
        self._input_prompt = simple_template or DEFAULT_SIMPLE_INPUT_PROMPT

    def _get_prompts(self) -> PromptDictType:
        """Get prompts."""
        return {"simple_template": self._input_prompt}

    def _update_prompts(self, prompts: PromptDictType) -> None:
        """Update prompts."""
        if "simple_template" in prompts:
            self._input_prompt = prompts["simple_template"]

    async def aget_response(
        self,
        query_str: str,
        text_chunks: Sequence[str],
        **response_kwargs: Any,
    ) -> RESPONSE_TEXT_TYPE:
        # NOTE: ignore text chunks and previous response
        del text_chunks

        if not self._streaming:
            return await self._llm.apredict(
                self._input_prompt,
                query_str=query_str,
                **response_kwargs,
            )
        else:
            return self._llm.stream(
                self._input_prompt,
                query_str=query_str,
                **response_kwargs,
            )

    def get_response(
        self,
        query_str: str,
        text_chunks: Sequence[str],
        **response_kwargs: Any,
    ) -> RESPONSE_TEXT_TYPE:
        # NOTE: ignore text chunks and previous response
        del text_chunks

        if not self._streaming:
            return self._llm.predict(
                self._input_prompt,
                query_str=query_str,
                **response_kwargs,
            )
        else:
            return self._llm.stream(
                self._input_prompt,
                query_str=query_str,
                **response_kwargs,
            )