|
Algorithm 1 Prompt generation process. |
-
1:
Input: Stories from input file
-
2:
Output: Processed stories with generated prompts
-
3:
Text Extraction:
-
4:
Extract key details from each story;
-
5:
Tokenization: Split text into words, remove stopwords;
-
6:
Frequency Analysis: Identify common words and themes;
-
7:
Character Extraction: Use regex to find capitalized names;
-
8:
Sentence Extraction: Extract first and last sentences;
-
9:
Prompt Generation:
-
10:
Use extracted elements to create prompts;
-
11:
Random Sampling: Select up to 5 themes and 3 character names;
-
12:
Template Usage: Apply various storytelling templates;
-
13:
Implementation:
-
14:
Read stories from input file (various text formats);
-
15:
Generate unique prompts for each story;
-
16:
Append prompts to story objects;
-
17:
Save updated stories to output file in JSON format;
-
18:
Create final dataset with titles and prompts in new JSON file;
|