Extract data from Companies House accounts in a streaming way.
Use this Python package to convert the zipped XBRL format that Companies House supplies into a single data frame of 38 columns.
Get startedMemory
Companies House accounts files can be large, and stream-read-xbrl only needs a small amount to be in memory at once.
Performance
On a standard laptop with 8 CPU cores it takes approximately 10 seconds to convert a single day of Companies House accounts data.
Contribute
The repository for stream-read-xbrl is public and contributions are welcome though GitHub.