DO NOT PLACE A BID WITHOUT READING THE DESCRIPTION TO THE END! IF YOU’RE GOING TO INCREASE YOUR OFFER IN MESSAGES AFTER BIDDING, THEN DON’T BID AT ALL! I have a completed Python-based project that uses PaddleOCR to extract data from a specific mobile application and export it to Excel. The system works well overall, but I need a few optimizations and new features: Panel Sub-Header Order Optimization: The main headers and Excel output are perfectly fine. However, sometimes the *sub-headers inside the panel* appear in the wrong order, which causes misalignment under their categories in Excel. The system should be optimized so that it **never disrupts the order seen in the mobile application panel**. Upgrade PaddleOCR 2.7 → 3.2: The project currently uses PaddleOCR 2.7. It should be upgraded to PaddleOCR 3.2 in a way that avoids dependency conflicts. I expect this to improve text accuracy. Numbers are always correct, but text recognition often fails. To handle this, I have already created a *corrections* section that can map misread values. New Data Extraction: An additional page in the mobile application contains 3 new data fields. These should be scraped and added as 3 new columns at the end of the Excel sheet. Other Minor Fixes: There may be a few small tweaks during development, but the above are the critical tasks. Delivery Time: 2–3 days maximum. Budget: Fixed within my stated range (please do not propose outside this range). In return, I will provide a 5-star rating and a positive review.