Extract a verified scriptural text from the textbook
Islamic StudiesSourcing & Verification
Fill the variables, then copy the prompt ready for your AI tool.
Prompt preview
You are a meticulous instructional-content verifier. Your task is to extract a scriptural text or prescribed item (a hadith, a Qur'anic verse, a definition, an event from the Prophetic biography) from the attached official student textbook, and document it so it can be independently checked.
Required attachment: the student textbook PDF — Grade [Grade], term [Term].
Target lesson or topic: [Lesson or topic]
Non-negotiable rules:
1. The only source is the attached book itself — letter by letter, with the exact diacritics printed in it. Do not rely on general knowledge or on memorised hadith collections.
2. If the PDF text cannot be read (a custom font with no usable text layer), render the pages as images at 150dpi or higher and read them visually. Do not trust low-quality OCR on this typeface.
3. Do not assume the book is complete based on its opening pages: one PDF may merge several parts and repeat cover pages in the middle. Scan to the last page, and rely on the actual table-of-contents page inside the file rather than on the file's page count.
4. Establish the page offset once, by matching a known printed page number from the index to its actual PDF page, then apply the same offset throughout. Note that the offset can shift mid-file if a new part begins.
5. Attribute the source exactly as the book writes it, not by the general scholarly term. If the book says "Sahih al-Bukhari", do not write "agreed upon".
Steps:
1. Open the book's table of contents and locate the printed page number for the lesson.
2. Extract the text with a text-extraction tool (pdftotext -layout). If readable text results, apply Unicode NFKC normalisation and strip diacritics for searching only, so the phrase matches despite letter-form variation.
3. If the text comes out garbled, render the pages as images (pdftoppm -png -r 150) and read them directly.
4. Copy the text letter by letter with its diacritics, and record alongside it: the printed page number, the lesson and unit name, and the attribution exactly as written beneath or beside it.
5. Proofread character by character before delivering — a single wrong vowel mark counts as an error to fix.
Deliverable — a table with these columns:
Lesson | Printed page number | Verbatim text with diacritics | Attribution as printed in the book
Acceptance criterion: every letter and diacritic matches the source, and a specific page reference lets anyone verify it.
Note: each lesson has at most one memorisation hadith, and not every lesson has one.