Document limitations of Bounter #37
Conversation
merge Rare into aneesh
|
Thanks @aneesh-joshi @piskvorky any comments? |
README.md
Outdated
| ```python | ||
| from bounter import bounter | ||
| bounts = bounter(size_mb=1) | ||
| bounts.update(str(i) for i in range(1000000)) |
There was a problem hiding this comment.
Add another 0 (10000000), to make it clear this is not related to 1 MB.
README.md
Outdated
|
|
||
|
|
||
| ## When not to use Bounter? | ||
| Beware, Bounter is only a probabilistic frequency counter and cannot be relied on for fine counting. (You can't expect a data structure with finite size to hold infinite data.) |
README.md
Outdated
| 0 | ||
| ``` | ||
|
|
||
| Please use `Counter` or `dict` when such fine counts matter. When they don't matter, like in most NLP applications with a huge corpora, Bounter is a very good alternative. |
README.md
Outdated
| 0 | ||
| ``` | ||
|
|
||
| Please use `Counter` or `dict` when such fine counts matter. When they don't matter, like in most NLP applications with a huge corpora, Bounter is a very good alternative. |
There was a problem hiding this comment.
…like in most NLP and ML applications with huge datasets, …
README.rst
Outdated
| When not to use Bounter? | ||
| ------------------------ | ||
|
|
||
| Beware, Bounter is only a probabilistic frequency counter and cannot be relied on for fine counting. (You can't expect a data structure with finite size to hold infinite data.) |
There was a problem hiding this comment.
Is this file auto-generated? Or why are we keeping two parallel READMEs?
Anyway, the same changes here please.
There was a problem hiding this comment.
In this case, I propose to stay only .rst (because this used for PyPI too) and drop markdown version (to don't support 2 almost same versions)
There was a problem hiding this comment.
I wonder if they're out of sync already… @aneesh-joshi can you check?
There was a problem hiding this comment.
@piskvorky
I will check. The .rst is not auto generated. I manually edited it.
|
Good idea! Thanks. |
|
@piskvorky README rst & md really not synced yet. I still think that we should stay only one version (create best .rst version & drop markdown version to avoid "de-sync" state). |
Did we already discuss this? To me a single version ( |
|
I don't remember (I'm even not sure if we discussed this) Maybe @isamaru remember? Right now I see no pros to maintain both |
|
@menshikh-iv @piskvorky Some questions:
|
README.md
Outdated
| bounts['100'] | ||
| 0 | ||
| ``` | ||
|
|
||
| Please use `Counter` or `dict` when such fine counts matter. When they don't matter, like in most NLP applications with a huge corpora, Bounter is a very good alternative. | ||
| Please use `Counter` or `dict` when such exact counts matter. When they don't matter, like in most NLP and ML applications with a huge datasets, Bounter is a very good alternative. |
There was a problem hiding this comment.
a huge datasets => huge datasets.
There was a problem hiding this comment.
Shoot! Missed that somehow.
README.rst
Outdated
| bounts['100'] | ||
| 0 | ||
|
|
||
| Please use ``Counter`` or ``dict`` when such fine counts matter. When they don't matter, like in most NLP applications with a huge corpora, Bounter is a very good alternative. | ||
| Please use ``Counter`` or ``dict`` when such exact counts matter. When they don't matter, like in most NLP and ML applications with a huge datasets, Bounter is a very good alternative. |
Yes
is http://pandoc.org/try/ works OK? |
|
@menshikh-iv but the generated file has a huge diff compared with the original .rst |
|
@aneesh-joshi that's expected: if you looking into commit history, you'll see a huge diff between this 2 files. Here you need to do some manual work ("join" both to one file). |
|
@menshikh-iv |
|
@aneesh-joshi okay, but please raise an issue about syncing |
|
@menshikh-iv |
|
ping @menshikh-iv @piskvorky |
|
Thanks @aneesh-joshi 🚀 |
menshikh-iv
changed the title
Credit goes to the people on this page.
I merely worded their concerns.
Fixes #36 .
CC: @menshikh-iv