Just finished my 3 AM coffee, trying to export some annotated PDF data from ChatPDF for a compliance report. The export function is, to put it kindly, having a bad night. Every CSV it spits out is corrupted. Opens in Excel? Gibberish. `cat` it in the terminal? Throws encoding errors.
Here’s what I’m seeing:
- Selecting "Export conversation" -> CSV.
- File downloads as `chatpdf_export.csv`.
- Immediate check with `file -I` shows `charset=unknown-8bit`.
- Attempting to parse it with a simple Python script throws a `UnicodeDecodeError`.
```python
import csv
with open('chatpdf_export.csv', 'r') as f:
reader = csv.reader(f)
for row in reader:
print(row)
```
Result: `UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0`
Tried different browsers (Chrome, Firefox) and different PDFs. Same result. It's like it's writing raw bytes with no proper encoding header.
Has anyone else hit this? Is there a hidden "export as UTF-8" toggle I'm missing, or is this a known bug? My workaround right now is copying the text manually, which feels like deploying a cluster with a shell script 😅
Pager duty survivor.
NightOps