Tools
Use this tool for more refined results
Use this tool to locate thematically similar texts
Use this tool to scan any 19th or early 20th century fiction text, or groups of fiction texts up to 100 at a time, (.txt files only) to detect 'science in fiction' according to historic definitions of vintage science fiction by the public at the time, i.e. science fiction should contain realistic science terminology. This tool scans for over 3000 science related words used in the 19th and early 20th century and gives a density score for each txt to indicate the probability of a text being science fiction. (Density calculation percentage based on an arbitrarily chosen average of 20 science words per 1000 words across the entire text signifying 100% probability the text is science fiction.) You can then export the results as a csv. Program plan researched and constructed by Neil Hogan with terminology help from Claude and final output vibe coded through Replit.
Automated Popular English Fiction Genre Identification Program
The Python code uses the Open API service so requires $$$ if you wish to use it. Read the code I used to determine the genres of 25,000 texts here. (It cost me about USD$1000 plus time, but I preferred that to sending the entire corpus to OpenAI for private processing at 50% off.)
Test the APEFGIP online with your own Open API key here.
Download a zipped version for your Python environment here (some code restrictions embedded which you can change later after testing)
Use this tool to create your own ebook containing texts from this database. Note that this service will give you any text, even those that have poor OCR. Choose the text you would like in your ebook, generate a basic cover, then output the collection as a pdf. (Delete anything in the Gemini key field first). Or enter your gemini key to generate your own cover. Note that due to artefacts and poor OCR, some sentences may not be able to wrap within the margins. In which case some judicious editing of your exported pdf will be required. Program plan researched and constructed by Neil Hogan with final output vibe coded through Replit.