WIP: Implement 'ipython history show [opts]' #13568
Conversation
Includes: * a `-l` / `--limit` option which defaults to the last 1000 entries from the SQLite history database * a `-P` / `--pretty` option which pretty-prints the output with IPython.util.PyColorize * flags to specify a custom delimiter (defaults to tab) and add a header row (corresponding to the SQLite column names) * flags which control `--pretty`'s colorized output, defaulting to color only when stdout is a terminal
|
Closing old stale PRs, sorry if this was not addressed, but this has been a few years, so there are few chances of someone actually pushing this through. Feel free to resubmit if you feel this has a chance to be revived. |
@Carreau Do you think it has more of a chance of being merged if I am simply more confident with the changes and the PR writeup (i.e., don't ask any questions, or ask for feedback)? I can see how that might be psychologically off-putting / burdensome to an already-spread-thin open source maintainer. If it is a desirable feature, I will rebase the changes against some recent branch, no problem. Otherwise I can simply maintain it as a local patchset that I apply to each new IPython installation manually. |
|
I mostly need time to review. I can try. one thing: I think you can use 'with conn' instead of manual closing . |
| show_hist_help = """Show recent entries from the IPython history database. | ||
|
|
||
| Limited to the last 1000 unless `--limit=N` is supplied; `--limit=0` or `-1` | ||
| prints all entries in the database. | ||
| """ | ||
|
|
There was a problem hiding this comment.
You can inline this directly if you wish, you don't have to rely on a seperate variable. Up to you.
| colors = CaselessStrEnum( | ||
| ("Neutral", "NoColor", "LightBG", "Linux"), | ||
| default_value="Neutral", | ||
| help="Set the color scheme (NoColor, Neutral, Linux, or LightBG).", | ||
| ).tag(config=True) |
There was a problem hiding this comment.
This have changed in 9.0, now we have themes, that are arbitrary strings, the Neutral, CoColor, LightBG, and Linux still exists but are lower case.
| def start(self): | ||
| profile_dir = Path(self.profile_dir.location) | ||
| hist_file = profile_dir / "history.sqlite" | ||
| con = sqlite3.connect(hist_file) |
There was a problem hiding this comment.
It hink we can do a with sqlite3.connect(hist_file) as con here.
| import sys | ||
| from IPython.utils import PyColorize |
There was a problem hiding this comment.
I think we can import those unconditionally at top of file.
| # FIXME: even if you specify `-d,` this is not very | ||
| # CSV-friendly, since the outer quote character (single or | ||
| # double) depends on whether the history entry itself has | ||
| # embedded strings | ||
| print(self.delimiter.join([repr(x) for x in entry])) |
There was a problem hiding this comment.
I think that's fine as is for now.
|
@Carreau thanks for the thoughtful feedback! This is on my radar, and I'll be sure to note what behavior has changed in recent releases before I update the PR. |
Use case: I want to quickly copy-paste some statements from the history of a recent IPython session into another source file.
Previously, I may have had some hacky shell function that involved running
sqlite3, but I'd always have to hunt for where the SQLite database was, and that was irritating. What I was looking for was something comparable tohistoryin the Unix shell; easy to pipe intogrep,pbcopyorxclip, et cetera.This PR includes:
-l/--limitoption which defaults to the last 1000 entries from the SQLite history database-P/--prettyoption which pretty-prints the output withIPython.util.PyColorize--pretty's colorized output, defaulting to color only when standard output is a terminalThings I would appreciate feedback on
show?display?print?tail? aliases for all four?--color=alwayswhen--prettywasn't givenlogging.warning-pis OK as an alias for--pretty, or if that's likely to conflict with some other optionhistoryapp.pypytestin the top-level repository directory, I just getexamples/IPython Kernel/ipython-completion.bash, but that can't be the "real" completion script if it's inexamples, can it?To-do
prdirectory