|
Description: |
Expected duration: 3 - 4 weeks
Requirements of the contractor:
- must be experienced in coding/programming. Specific Arabic language libraries is highly desirable
- must be Arabic speaking in and able to review Arabic Fus'Ha including tashkeel
Requirements of the project: All of the below must be met for payment since there are already online tools much cheaper than I'm paying here that do some of the below but not all.
For full payment I can accept up to 5 errors a page
For half payment I can accept up to 15 errors per page
NOTE: where I mention Python or libraries below, this are just example. You can use any method you deem fit, as long as the product is delivered.
Below are the original Arabic work and the English translation. You can start from the Authors preface. The footnotes would be different so use google translate to translate the footnotes not the English translation.
Original Arabic word to be converted and translated: https://archive.org/details/AlIrshad_201712/%D8%A7%D9%84%D8%A5%D8%B1%D8%B4%D8%A7%D8%AF%20-%20%D8%A7%D9%84%D8%B4%D9%8A%D8%AE%20%D8%A7%D9%84%D9%85%D9%81%D9%8A%D8%AF%201/page/n3/mode/2up
English Translation: https://archive.org/details/AlIrshad_201712/%D8%A7%D9%84%D8%A5%D8%B1%D8%B4%D8%A7%D8%AF%20-%20%D8%A7%D9%84%D8%B4%D9%8A%D8%AE%20%D8%A7%D9%84%D9%85%D9%81%D9%8A%D8%AF%201/page/n23/mode/2up
All the below must be delivered:
1. Convert PDF to Clean Word Document: online tools have some errors when converting this so you would need to use coding to ensure accuracy, as standard converter tools have failed to maintain letter integrity, often replacing letters with shapes or symbols.
2. Font and Size:
English text: Arial, size 11.
Arabic text: Arial, size 15.5.
3. Document Split: Split the final output into separate Word documents each one having finishing on the 10th page e.g. first document 1-10, second document 11-20, third 21-30 and so on. This is instead of one large document which would make it slow to work on Google docs
4. Create a Two-Column Table: Place the original Arabic text in the right column and the English translation in the left column.
5. Alignment: Ensure Arabic text is right-to-left aligned.
7. Add Tashkeel to the Arabic Text: for example utilize a Python library or function to add tashkeel to the Arabic text.
8. Intelligent Sentence Splitting: Split the text into 1 or 2 sentences per cell using an Arabic-specific NLP library rather than splitting by commas or fullstops. For example libraries like camel-tools can tokenize Arabic sentences while respecting linguistic boundaries (e.g., keeping “bihaa” with its associated words).
9. Sentence matching: Ensure Arabic sentences of each cell are matching the correct English translation sentence. This will require some work since the English translation may either omit a sentence completely or give a completely different meaning than the literal meaning. Where completely omitted the English cell can state "sentence omitted" or meaning is altered significantly this can stay in the cell but a comment explain this.
10. Quality Expectations: The majority of the tashkeel, sentence splitting, and translation must be accurate, with no more errors than what is specified above.
Trial the process on 1-3 pages first and submit for quality check before proceeding further.
The final output must maintain high quality across all pages.
Random checks will be conducted to ensure quality. |
