@admin_mike, I watched your video explaining how a person can upload a CSV file of data, and PickAxe will treat each row as a separate chunk.
Is there a way that a CSV can be used to upload TEXT chunks that have been optimized to avoid boundary issues and ensure each chunk make sense on its own? Each “row” would have just one column and would contain one optimized chunk of text within a recommended size range.
I’ve tried this, using Excel saving to CSV, and also manually creating a quote (") delimited CSV, both without success. PickAxe finds relevant data in a standard txt file that it has chunked automatically, but it doesn’t seem to find the same text in a CSV file created with Excel or manually. PickAxe accurately reports the number of chunks I’ve created in my CSV file, but it doesn’t seem to be able to use that data.
I’m looking for a way to bulk update an optimized KB file without losing existing optimizations and without having to do chunk-by-chunk editing in Pickaxe.
Yes. If you put lots of text into a spreadsheet, you can upload it into the Knowledge Base. I often favor this format because it gives me so much control. For example, in the Doorman Pickaxe I have a spreadsheet with a column for user question and a column for the answers I’ve written. Uploaded into the KB, each row is added as an individual chunk, so I ensure that each chunk has both the full question and the full answer when it’s recalled.
Follow-up questions:
– What do you consider the optimal chunk size, and also the maximum chunk size?
– In your use case, do you upload it as a spreadsheet file (.xls), or a .csv?
– Do newline symbols (\n) cause any issues? (When I experimented briefly last week with uploading text in a .csv, the text showed up in the chunk Explorer with many newline characters.)