Perform granular updates if reloading textual files #366
Conversation
jupyter_ydoc/yunicode.py
Outdated
|
|
||
| # for very different strings, just replace the whole content; | ||
| # this avoids generating a huge number of operations | ||
| if matcher.ratio() < 0.6: |
There was a problem hiding this comment.
Using 0.6 because:
As a rule of thumb, a ratio() value over 0.6 means the sequences are close matches
As per https://docs.python.org/3/library/difflib.html#sequencematcher-examples
Instead of ratio() we could use quick_ratio() or real_quick_ratio(). In case if the ratio is above the threshold this does not matter because we would end up computing matching_blocks anyways (this is computed and cached by a call to either raito() or get_opcodes(); in that case calling quick_ratio heuristic could actually end up taking more time as we end up doing more work. It all boils down to whether we think that we are more likely to see smaller updates or larger updates. One very cheap heuristic could be using len of both strings which is what real_quick_ratio does.
1 participant
Uh oh!
There was an error while loading. Please reload this page.