Add documentation for spellchecker and spellcheck docs #2025
Conversation
|
Note: I briefly messed around with the spellchecker, and wasn't immediately able to get it working on Linux. It looks like the script as it currently stands is Mac-only, but even manually installing Aspell, I couldn't get it to use |
Thanks @zmbc for drawing this to our attention! Will have a look into it :) |
Thanks again @zmbc. Can I just check which Linux distribution and version you're running? I'm assuming you tried to run the spellchecker following the instructions in this PR - is that correct? If so, it is probably not working because extra dependencies are required for Linux, e.g. we've been able to run the spellchecker on a Docker file using alpine 3.14 but this required installing some additional dependencies: |
|
@zslade Sorry -- I should have given a bit more information. I'm on Ubuntu 20.04 via Windows Subsystem for Linux. I installed aspell via conda. I can run the spellchecker script (if I comment out the Mac-specific part) but I have no reason to believe that it is using the LibreOffice dictionaries. |
|
Running on |
Thanks @zmbc. Just so I'm really clear - the spellchecker runs for you (spellchecks the docs) but doesn't pass (finds spelling errors)? We are still in the process of updating As a rough check, could you please Would you also be able to tell us which Mac-specific parts you commented out? This will help us if we decide built a system-agnostic version :) Many thanks! |
I get
Correct, and a lot of them, so I figured that wasn't working correctly. And, I'm pretty sure it's not using the
I commented out these lines, since they use Homebrew, which is Mac-specific: splink/scripts/pyspelling/spellchecker.sh Lines 13 to 15 in 1d942f0 I replaced them with conda package installation (aspell and go-yq), which is cross-platform. Happy to contribute this if you are interested. Would you be open to switching to that, or would you want to have both?
Ah okay, I think this was the crux of my misunderstanding. I was working on doc updates in #2083 and thought I needed to get this spellcheck passing. The name of this PR indicated to me that the feature was done, just not documented. |
|
@ThomasHepworth, I think all mistakes have been corrected now 🙌 . See the latest diff for the decisions I made related to your PR #2102.
|
Type of PR
Is your Pull Request linked to an existing Issue or Pull Request?
Docs spellchecker update: #2000
Original spellchecker PR: #1588
Give a brief description for the solution you have provided
I have updated the documentation with guidance on how to use the docs spellchecker, following on from this issue #2000.
All docs have been spellchecked to set a baseline for 'spellchecked' docs. This involves
custom_dictonary.txtwhich aren't being picked up by LibreOffice dictionaries, e.g. words like Splinkpyspelling.ymlto bypass certain patterns or sections in the docs, e.g. code blocks@ThomasHepworth and I decided not to make the spellchecker part of CD/CI with a GitHub Action at this stage as would require some non-trivial configuration and a new check box on the PR template will probably suffice.
Code for testing ignore rules in pyspelling.yml
General text
This is some text that does not contain a mistake
startoffilemistake
colons
With a massive thanks to external contributor @hanslemm, Splink now supports :simple-postgresql: Postgres. To get started, check out the Postgres Topic Guide.
[:colon-mistake: RSS feed]
:octicons-duplicate-24-mistake:
Code block ignore test
python code
pycodeblockmistakeeealt python blocks
inline code
inlinecodemistkakeeeel.first_name = r.first_name and substr(l.surname,1,1) = substr(r.surname,1,1).blog post headers
date: 2024-01-23
authors:
categories:
anchor tags
anchormistake
Test text between
diffblocks??? info "Example of convergence output"
links
create_functionfunctiontext between
$$$$\text{}
mathcodeblockmistake
\something $$
Almost 100%, say 98%$inlinemathmistake m \approx 0.98$
compound words
Equi-join
Card content ignore test
::cards::
[
{
cardblockmistakee
},
]
::/cards::
auto generated code
::: splink.linker.Linker
handler: python
selection:
members:
- cluster_pairwise_predictions_at_threshold
- compare_two_records
- compute_tf_table
- deterministic_link
- find_matches_to_new_records
- load_settings
- load_model
- load_settings_from_json
- predict
rendering:
show_root_heading: false
show_source: true
BibTeX blocks
error logger
To enable the logging of multiple errors in a singular check, or across multiple checks, an
ErrorLoggerclass is available for use.The
ErrorLoggeroperates in a similar way to working with a list, allowing you to add additional errors using theappendmethod. Once you've logged all of your errors, you can raise them with theraise_and_log_all_errorsmethod.??? note "
ErrorLoggerin practice"```py
from splink.exceptions import ErrorLogger
endoffilemistake
PR Checklist