Please understand this story and provide all the characters and key words (gender (can be supplemented with imagination, but must be clear), age (can be supplemented with imagination, but must be clear), race (can be supplemented with imagination, but must be clear),
Please understand this story and provide all the characters in it, with multiple characters separated by commas`;
skin color (can be supplemented with imagination, but must be clear), clothing (can be supplemented with imagination, but must be clear), hairstyle (can be supplemented with imagination, but must be clear), hair color (can be supplemented with imagination, but must be clear),
facial features (can be supplemented with imagination, but must be clear). \n
roles=roles.replace(/。/g,'').replace(/、/g,',')
console.log(roles)
constroles_arr=roles.split(/[，,]/);
console.log(roles_arr)
asyncfunctionprocessRoles(){
for(constone_roleofroles_arr){
awaitdelay(100);
constadapt_keyword_restrict=`
Instructions:
Please understand this story and provide the keywords for the character "${one_role.trim()}" (gender (can be supplemented with imagination, but must have it), age (can be supplemented with imagination, but must have it),
Skin color (can be supplemented with imagination, but must have it), clothing (can be supplemented with imagination, but must have it), hairstyle (can be supplemented with imagination, but must have it),
Hair color (can be supplemented with imagination, but must have it), facial color (can be supplemented with imagination, but must have it), facial features (can be supplemented with imagination, but must have it).
Requirement:
Requirement:
The roles and keywords should correspond.
Keywords are separated by commas.
Strictly return in the following format: [{"Role": "", "Role Keywords": ""}]`;
As long as the keyword is returned, no additional explanatory text is required.`;
Please understand this story and provide the following description for the scene: "${item.场景描述}":
Please understand this story and provide the keywords for the scene "${item.场景描述}" (era (can be supplemented with imagination, but must be present), space (can be supplemented with imagination, but must be present),
scene keywords (era, space, time period, geographical environment, weather, objects, characters, camera angle)
Time period (imagination can be used to supplement, but it must be present), geographical environment (imagination can be used to supplement, but it must be present), weather (imagination can be used to supplement, but it must be present),
characters (select the character in this scene from all characters)
Items (can be supplemented with imagination, but must be present), characters (can be supplemented with imagination, but must be present), camera angles (can be supplemented with imagination, but must be present).
character keywords (select the character keywords for this scene from all characters and copy them as they are).
Requirement:
Requirement:
The roles and keywords should correspond.
Keywords are separated by commas.
Strictly return in the following format: [{"Scene Keywords": "(Multiple words separated by commas)", "Role": "(Characters separated by commas)", "Role Keywords": "(Give corresponding character keywords separately, multiple words separated by commas)"}]
As long as the keyword is returned, no additional explanatory text is required.`;
Do not add any other instructions except for the content returned in format.
Please understand this story and for the scene: "${item.场景描述}", select the character in this scene from all the characters, with multiple characters separated by commas.`;
constitem_roles=awaittext2videoService.submitGpt("story:\n"+form.chatgpt_answer+"\n all the characters: \n"+JSON.stringify(form.chatgpt_answer_roles)+"\n"+adapt_role_restrict,tuili_llm);
