You've already forked bilibili-subtitle
问题生成
This commit is contained in:
@@ -8,6 +8,7 @@ export const STORAGE_TEMP = 'bilibili-subtitle_temp'
|
||||
export const PROMPT_TYPE_TRANSLATE = 'translate'
|
||||
export const PROMPT_TYPE_SUMMARIZE_OVERVIEW = 'summarize_overview'
|
||||
export const PROMPT_TYPE_SUMMARIZE_KEYPOINT = 'summarize_keypoint'
|
||||
export const PROMPT_TYPE_SUMMARIZE_QUESTION = 'summarize_question'
|
||||
export const PROMPT_TYPE_SUMMARIZE_BRIEF = 'summarize_brief'
|
||||
export const PROMPT_TYPE_ASK = 'ask'
|
||||
export const PROMPT_TYPES = [{
|
||||
@@ -22,6 +23,9 @@ export const PROMPT_TYPES = [{
|
||||
}, {
|
||||
name: '总结',
|
||||
type: PROMPT_TYPE_SUMMARIZE_BRIEF,
|
||||
}, {
|
||||
name: '问题',
|
||||
type: PROMPT_TYPE_SUMMARIZE_QUESTION,
|
||||
}, {
|
||||
name: '提问',
|
||||
type: PROMPT_TYPE_ASK,
|
||||
@@ -46,6 +50,12 @@ export const SUMMARIZE_TYPES = {
|
||||
downloadName: '💡视频要点💡',
|
||||
promptType: PROMPT_TYPE_SUMMARIZE_KEYPOINT,
|
||||
},
|
||||
question: {
|
||||
name: '问题',
|
||||
desc: '常见问题',
|
||||
downloadName: '💡常见问题💡',
|
||||
promptType: PROMPT_TYPE_SUMMARIZE_QUESTION,
|
||||
},
|
||||
}
|
||||
|
||||
export const PROMPT_DEFAULTS = {
|
||||
@@ -124,6 +134,46 @@ The video's subtitles:
|
||||
'''
|
||||
{{segment}}
|
||||
'''`,
|
||||
[PROMPT_TYPE_SUMMARIZE_QUESTION]: `You are a helpful assistant that skilled at extracting questions from video subtitle.
|
||||
|
||||
## Context
|
||||
|
||||
The video's title: '''{{title}}'''.
|
||||
The video's subtitles:
|
||||
|
||||
'''
|
||||
{{segment}}
|
||||
'''
|
||||
|
||||
## Command
|
||||
|
||||
Accurately extract key questions and their corresponding answers from the video subtitles based on the actual content provided. The number of questions should be between 3 and 5.
|
||||
|
||||
- Identify questions as sentences starting with interrogative words (e.g., "What", "How", "Why") and extract the following sentences that directly answer these questions.
|
||||
- Include only those questions and answers that are relevant to the main points of the video, and ensure they cover different aspects of the video's content.
|
||||
- If an answer spans multiple non-consecutive parts of the subtitles, concatenate them into a coherent response without adding any information not present in the subtitles.
|
||||
- In cases where the number of potential Q&As exceeds 5, prioritize the most informative and directly answered ones.
|
||||
- If clear questions and answers are not available in the subtitles, refrain from creating them and instead note the absence of direct Q&As.
|
||||
- Answer in language '{{language}}'.
|
||||
- Format the output in markdown json format, as specified.
|
||||
|
||||
## Output format
|
||||
|
||||
Provide an example to illustrate the expected output:
|
||||
|
||||
\`\`\`json
|
||||
[
|
||||
{
|
||||
"q": "What is the main theme of the video?",
|
||||
"a": "The main theme of the video is explained as..."
|
||||
},
|
||||
{
|
||||
"q": "How is the topic developed?",
|
||||
"a": "The topic is developed through various examples, including..."
|
||||
}
|
||||
]
|
||||
\`\`\`
|
||||
`,
|
||||
[PROMPT_TYPE_ASK]: `You are a helpful assistant who answers question related to video subtitles.
|
||||
Answer in language '{{language}}'.
|
||||
|
||||
|
Reference in New Issue
Block a user