NumPy Documentation Team meetings

--- tags: NumPy --- # NumPy Documentation Team meetings - Time: 2 pm UTC - Join via Zoom: https://us06web.zoom.us/j/85016474448?pwd=TWEvaWJ1SklyVEpwNXUrcHV1YmFJQT09 (To dial in, find your local number: https://us06web.zoom.us/u/ksv5WeoUL) - NumPy community events calendar: https://scientific-python.org/calendars/ - Next meeting: September 26, 2022 - [Documentation brainstorming document](https://hackmd.io/RdtnQZpLRZqgNRe4gaJ0SA) - anything that is not yet ready for an issue or PR or needs a discussion ### Code of Conduct We want to take a moment to remind you that this meeting, like all project spaces, is meant to be open, welcoming, diverse, inclusive, and it's important for us to have a healthy community. Like all NumPy spaces, and everyone participating in them, this meeting will follow our [Code of Conduct](https://numpy.org/code-of-conduct/). If you haven't read it yet, please take some time to do so later on as it already applies to you. For now, in short, please be kind and generous towards one another. ## Topics for September 26, 2022 ## Topics for September 12, 2022 **Present:** Melissa Mendonça (@melissawm), Inessa Pawson, Mars Lee (@marsbarlee), Sebastian Berg, Ryan Cooper, Matthew Murray - [name=Mars] Rough 1st draft of 'NumPy Contributor Adventure' - ![](https://i.imgur.com/EG3Z4BR.png) - Here’s a sketch I made of the discussion of contributor → maintainer, from the NumPy community call on August 31, 2022. - Here are meeting notes [on this conversation, in the NumPy Github repo](https://github.com/numpy/archive/blob/main/community_meetings/community-2022-08-31.md#new-topics). - I wrote the script. I tried to address some concerns I’ve heard about illustrating or codifying an informally defined path. - That’s you can’t skip ahead, it’s moderated but also very community-focused. Parts such as “with mentorship”, “not a quick or easy path, but one worth having” and “community will be there with you every step of the way” - Also [discussed in Open Source Design forum, got input from designers](https://discourse.opensourcedesign.net/t/open-source-comics-resources-collaboration-and-funding-for-making-comics-for-open-source-projects/3130/4) - Questions to call attendees - Tone - Light hearted - Accuracy - Can shift, but this journey is not formalized anywhere - Where in the world/documentation/internet could something like this exist? - A Contributor Experience handbook? Not just for NumPy? - Expanding on and future possibilities - Sharing illustration on Github - Write image description, people can comment on as Github suggestion - "A person holding up a sword" - Github suggestion higlighting this "A person holding a magic wand" - Similar process to existing design feedback process, but instead of linking out to external site and requiring account like [Figma](https://www.figma.com/file/7cVUnr5f4EG6yCRn2GJVQX/Untitled?node-id=0%3A1) - ![](https://i.imgur.com/ZzGtp9i.png) - Async vs sync process - Discussion notes - Melissa likes lighter tone for inviting community - Understand concern of codifying informally defined process, can be refined, not too rigid. Depends if we want to make specific for NumPy or include other projects and their workflows - Here, you need to ask for commit rights or fulfill minimum listed in project documentation eg 5 PRs - "Getting to Know": tag projects - Covers code tracks, doesn't cover non-code - eg contributor experience lead, design, social media, add extra branches - Maybe zoom out diagram before zooming in - Other possible roles - Steering council, Maintainer, Active contributor, Drive-by contributor, Experienced user, New user - Having both specific and broad versions - Scope out, refine NumPy version first - Generalize later - Inviting people for design feedback process - Melissa: Concerned that this puts more workload on designer to explain images - Is there a way share weight of work? - Granularity of image description - Puts depth of design process, educational - Inessa, have experience reviewing design work, can comment - Worth trying, current tools are all imperfect eg Online Adobe PDF review - SVGs are easier to version control - SVGs are mathematical formula, change code from 2 pixel width to 4 pixel width, hex code #ABC123 to hex code #ABC129 - But Github doesn't show SVGs in Github comments or Github-flavored markdown previews - Ryan - Also struggled to track changes in curriculum, diagrams - [Github for Poets video](https://www.youtube.com/watch?v=BCQHnlnPusY), designers using Github tracking of plain text - Broader discussion - Tracking non-plain text changes - Assimilationist approach: outsider takes on insider culture vs keeping distinct - designer taking additional responsibilities on developer workflows, platforms, Git and Github vs linking to Figma, designer workflow and platforms - Accomdating images on text platform: hitting storage limits since images are bigger than plain text, comparing differences - Next actions - [ ] [name=Mars] A draft PR, on Mars fork, on numpy.org repo - Not final submission place, but to start this design process experiment - Could go under https://numpy.org/contribute/ - A video recording of Sebastian Berg’s presentation “Ufuncs and dtypes: new possibilities in NumPy” has been posted on the NumPy YouTube channel: https://youtu.be/u9qU6cy5JkE. For the slide deck, visit: https://github.com/numpy/archive/blob/main/content/UFunc%2BDType_SciPy2022_presentation.pdf - [name=Melissa] [Documentation issues](https://github.com/numpy/numpy/labels/04%20-%20Documentation) ## Topics for August 29, 2022 **Present:** Melissa Mendonça (@melissawm), Inessa Pawson, Ross Barnowski, Mars Lee (@marsbarlee) - NumPy was selected as a featured project for [Grace Hopper Celebration Open Source Day 2022]( https://ghc.anitab.org/programs-and-awards/open-source-day/) held on Friday, September 16th, 8:00 am - 3:00 pm Pacific Time. We are currentlylooking for mentors who would be helping NumPy newcomers during this event. If you are available, please sign up here: https://docs.google.com/spreadsheets/d/1Y1t7-Yu2INR-Unl277TCJmlKOSNWok8fX2NbES5jcgk/edit?usp=sharing. You may choose to help for 2 hours only or stay longer. ***Who can be a NumPy mentor at a mentored/newcomers sprint?*** If you are comfortable providing guidance on git, NumPy development workflow, building NumPy documentation, setting up and using development environment, and using Gitpod, you are a great candidate for being a NumPy mentor to newcomers. - [name=inessa] DOC: Improve the description of the dtype parameter in numpy.array docstring: https://github.com/numpy/numpy/issues/10614 - Problem: Higly depends on individual set-up, would take too long to explain for each one - Possible solution: 'Check documention for d-type for details' - Does the documentation explain d-types sufficiently for this use case? - Link to d-type alias page? - [name=melissa] Comment on issue with summary of this discussion - Leave issue open for interested contributor or team could take over later - [name=inessa] Presentation topics for NumPy Newcomers’ Hour To suggest a topic for future presentations, please leave a comment in this tracking issue: https://github.com/numpy/archive/issues/58. - [name=Melissa] [Documentation user stories](https://github.com/numpy/numpy/issues/22089) - Feel free to add yours! - [name=Melissa] [Documentation issues](https://github.com/numpy/numpy/labels/04%20-%20Documentation) - [name=rossbar] numpy-tutorials: getting rid of gym/atari dependency * Licensing issues for end-user * Proposal, short-term fix: Make RL tutorial no longer executable * Long term fix: Someone to review issues, licensing, find a middle ground of execution without licensing trouble * But takes time, instead of letting this block us from solving licensing issue, we can do short-term fix * Another solution: Timestamp with version for articles about this, such as "This may not execute anymore" - [name=rossbar] Related to the above - what to do about non-executable tutorials (RL + NLP)? - Related discussion: sphinx-design - [name=Mars] Update on Small Development Grant proposal, [NumPy Comics](https://docs.google.com/document/d/14O9Z368nhZ8Dc6pd37XbDa5tGa86_zLlcgcCotbUw7Y/edit?usp=sharing): - May not move foward with grant proposal due to conflict of interest - Proposal artist already works for NumPy through employer (Quansight) - Options for work to move foward - Artist to do work as part of Quansight-NumPy hours instead of grant? - Send same proposal details to employer, get approval through employer? - Timeline: Q4 2022? 2023? - Future Quansight 2023 internship - Cross-promote to [Open Source Design internship commununity](https://discourse.opensourcedesign.net/t/continuing-conversations-about-fellowships-internships-in-open-source-design/2722/13)? - After competing this version as a 'pilot' - Discussion about compensation for other roles: some roles are done as community contribution not financial compensation - Other roles: 'Domain expert' to check comic accuracy - focus groups and user research - Combine Artistic world, Review world: differing platforms and workflows - Similiar problem to video-production, vidoe format does not easily - Show coming from NumPy, not just individual artistic work - SciPy comics, OSD call talked about getting disabled community to review alt text - Admit did not with conference timeline. One way to ensure include disabled communitity is compensation for review, which could come from a grant vs employer only compensate artist - Current proposal artist workflow: Sticking to Github workflow - Github is moving to improve but no clear answer yet - Simlar disucssion happening in [OSD forum: Open Source Comics: Resources, Collaboration, and Funding for making Comics for Open Source Projects](https://discourse.opensourcedesign.net/t/open-source-comics-resources-collaboration-and-funding-for-making-comics-for-open-source-projects/3130): - For next OSD community call, talk about incorporation of open source review process - This work can continue internally but continue to make process external - Next steps: - [ ] [name=Mars] Follow up with manager, employer - [ ] Bring this conversation (NumPy community, developer POV) back to OSD - Acccess to the plausible.io metrics Inessa will contact Stefan ## Notes for August 15, 2022 **Present:** Melissa (@melissawm), Inessa Pawson (@inessapawson), Mars Lee (@marsbarlee), Matti Picus (@mattip), James Munroe (@jmunroe) - [name=Inessa] NumPy documentation user stories - tracking issue: https://github.com/numpy/numpy/issues/22089 - [name=Inessa] NumPy is now on LinkedIn: https://www.linkedin.com/company/numpy/. Please follow us! - The communications policy: announcements only. - [name=Melissa] [DOC: Reorganize user guide outline #22117](https://github.com/numpy/numpy/pull/22117) - [Built artifact](https://output.circle-artifacts.com/output/job/31dcda2f-25a4-46d3-ba7f-0ce2c452865f/artifacts/0/doc/build/html/user/index.html) - Moves a couple of pages, while not exactly changing their content. - Open to suggestions for section names - Renamed 'basics' to 'fundamentals' - 'Under the Hood' or 'For Developers' or 'For Advanced Users'? - [name=Melissa] [Move to sphinx-design](https://github.com/numpy/numpy/pull/22116) - Currently use 'sphinx-panels', allows for cards - sphinx-panels is being deprecated - Move over sphinx-design before sphinx-panels is completely deprecated - Advantages of sphinx-design: more functionality, combines features of sphinx-panels, buttons, etc - Melissa to investigate (thanks @rossbar!), ensure CI passes so we can merge - [name=Melissa] Implement [Plausible tracking](https://plausible.io/) on all pages, not just NumPy.org - Tracks every doc page - bounce rate, where they come to the page, etc - Doesn't track personal data, better fit than Google Analytics - We should wait for the next release of the scientific python theme. Pamphile Roy mentioned it will include a configuration option similar to what is done to setup google analytics, so will be easier to set up - [name=Inessa] Updating content of [SciPy Lectures](https://scipy-lectures.org/) - James Munroe, Product Lead and Community Manager at 2i2c, would like to help with updating the content of the Lectures or some other topics related to the NumPy documentation - **Seeing missing piece: After newcomers complete the first tutorial,they think 'why continue with this tool?'. How does this replace my current tool?** - Convince people, show it can maximize utility - Stages of SciPy Lecture updates - 1. Infastructure - 2. Content - Does this work need to be done? - If not, we could just point SciPy Lectures as is, amplify - How will this work be done? - First step: Inessa to contact original creators to get permission to work on the content review, perhaps some of them would like to collaborate. - User groups James is focusing on - User: No problem installing tool, got buy-in, got small team of 10 people - "I did a 2 hour course, what else does this cool thing do?" - "We have a matlab license already, what does this do better?" - Formats aiming to make: Written, Lecture, Interactive course - Problem: not aware of open source toolset to make an interactive course, spin up enviroment for users to play in - One option (nascent): PyScript, running content in browser itself, - Grading? - Automatic grading tools such as [NB grader](https://nbgrader.readthedocs.io/) - [name=Ryan] As educator of NumPy at UConn, alternates between using and not using grading systems. Prefers giving individual feedback instead of nitpicky, automatic feedback - Striking balance of having automatic and personal teaching - In-person workshops at some point? - Providing training by having really strong documentation - Nature of stack: Always changing, hits snags if relying on 3, 4, 5 years old lectures - Eg. Having newcomer read page on python2 vs python3: Do I have to worry about that? Yes, but not anymore - Matter of keeping content up to date and then enthusiastically recommending - Why open source in particular? - [name=James] Build upon other initatives: [NASA's TOPS (Towards Open Science)](https://science.nasa.gov/open-science/transform-to-open-science), 6 years, 20,000 scientists, developing curricla, open publishing, open data - Try to make 2023 the 'Year of Open Source' - Thus there already is a strong value proposition for open science - Moving forward with [ENH: ndarray.__format__ implementation for numeric dtypes](https://github.com/numpy/numpy/pull/19550)? - not merged yet: no release notes, what happens with object arrays, etc - [name=Matti] We need people to push their own PR, carry the flag - [name=Meekail] Changing view from 'here's my work, please look at it when you can' -> Interactive process, need to push and pull - Maybe that's why people haven't carried the flag, maybe don't know, need do a few more PRs to understand process - Or have questions don't know how to ask "Do you need help moving forward" or "Join the meeting" ## Notes for August 1, 2022 **Present:** Melissa Mendonça (@melissawm), Inessa Pawson (@inessapawson), Eric Roberts, Meekail Zain (@micky774), Mars Lee (@marsbarlee), Ryan C. Cooper (@cooperrc) Follow-up from last meeting: - [x] [name=Melissa] **Make an issue with a top level outline of user guide** - [ ] [DOC: Reorganize user guide outline #22117](https://github.com/numpy/numpy/pull/22117) - [x] [name=Melissa] **Make an issue another about user journeys** - [ ] [name=Melissa] **Make a PR about plausible tracking on full docs** - Waiting on release - [ ] [name=Melissa] **Engage the mailing list for opinions** on #20645 * DOC: Add images illustrating how-to-io.rst: https://github.com/numpy/numpy/pull/17556 - a new contributor is welcome. * Related issue https://github.com/numpy/numpy/issues/17517 * The diagram would be a valuable addition to the NumPy documentation. Reproduce it in `.md` using Mermaid (https://mermaid-js.github.io/mermaid/#/) as it would be more beneficial from the maintenance and reproducibility point of view. * Discussion and comparison with [SciPy Contributor Quickstart](https://docs.scipy.org/doc/scipy/dev/dev_quickstart.html) * Would something like this be helpful? * [name=Meekail] Content is there. What would be useful is explaining why things are they are the way they are * Not need to recompile X * Make explicit workflows that more experienced NumPy users implicitly know * [name=Meekail] Comparison to another project: When working to make scikit-learn [NEP 50 compatible,](https://numpy.org/neps/nep-0050-scalar-promotion.html) having multiple builds was very useful * Useful if knew that workflow explicitly, ahead of time * Meekail has [running list of issues we want to submit as issues on GitHub but not ready yet](https://hackmd.io/@micky774/Hk2VZPSTc/edit) * **Compiled from many NumPy docs meeting notes, so this easier to keep track in one document** * **"We talked about this before, but not acted on. Let's read up on the notes and continue the discussion from there"** * **Instead of starting discussion from scratch again** * **We have visilibity of past issues but not visibliity of discussion** * [name=Meekail] This is not high prioity list to fill out. Not a complete list either. * If you know what changed, throw in this running list * Switching NumPy to Meson build system * Meekail thinking of writing SIMD documentation, talked to Matti, SIMD also discussed in SciPy * Meekail can document feedback on SIMD, send to Ralf * **While Meson not built for SIMD support, Meson developers are supportive of NumPy's input** * [name=Melissa] SciPy build comparison: 20 minute build now takes 5 minutes * Unfortunately doesn't improve doc build but still better * No description of [many NumPy labels](https://github.com/numpy/numpy/labels) * Inessa already identified duplicate labels * Some labels not clearly signalled for contributors * For example, when a maintainer add a label, the contributor not know what it means * Where are/Are there NumPy Governance documents? * [Current governance docs](https://numpy.org/devdocs/dev/governance/index.html) * [Current NumPy team](https://numpy.org/teams/) * **No clear contributor -> maintainer flow right now** * Discussed before in the [NumPy CZI DEI grant meetings](https://chanzuckerberg.com/eoss/proposals/advancing-an-inclusive-culture-in-the-scientific-python-ecosystem/) * [name=Inessa] **Great question that's not asked enough, Meekail, let's ask again in the wider NumPy Community meeting** * Difficulty knowing who has what rights * "Wait, you don't already have triage rights?" * Comparison to other projects such as scikit-learn processes: More structured/bureaucratic but thus more clear flow * NumPy can take inspiration * How can we systemically talk about contributors? -> a [good possible Scientific Python Ecosystem Coordination (SPEC) topic](https://scientific-python.org/specs/) * SPEC: Macroscopic discussion board for cross-community projects * [name=Inessa] **For anyone: If you would like triage rights, just ask. We'll likely say yes. We need help triaging issues.** * How can we more clearly signal in the right places we need this kind of help? * **How can we clearly signal "Ask for triage rights"?** * Streaming, recording discussions about NumPy * Brought up in previous NumPy Docs meetings * Meekail, Matti had a great and insightful discussion on SIMD... but regret not recording all this valuable information * Automatically offering Meekail and Matti discussion, should we extend systemitically to other people and topics? * [name=Inessa] Great NumPy Newcomer's Meeting topic! * [ ] [name=Meekail] Talk to Matti to check, then will coordinate with Inessa to set-up, record and upload to NumPy's Youtube channel * How do we decide on NumPy Newcomer's Meeting topics? * Put out poll in NumPy #newcomers channel, get engagement? * Currently we are going down Inessa's idea list, but already running low, so polling for more topics would greatly appreciated * Preferred polling method? Discussion on community engagement methods * Freeform suggestions has least enagagement: Newcomers don't know what they don't know * The simpler the better? * Make a call at end of Newcomer's Meetings, ask audience to vote for next topic? * Meekail's 3 topics he would love to talk about for Newcomers * SIMD * Array Objects and Array Methods * Multi-Array models (how NumPy handles assembled modules) * [Do we have a place for community activity disucssion?](https://github.com/numpy/archive) * Preserve and track the dicussions from this meeting about NumPy Newcomer topics, polling, Youtube content, Youtube transcriptions * Yes, currently in the [Archive repo](https://github.com/numpy/archive) * Github might be a better place than NumPy Slack, more open than invite-only Slack * Github Discussions? * [name=Ryan] Has good experience with Github Discussion in classroom * Some people scared this would turn into a user dicussion forum, different and more workload * A community repo? * [ ] [name=Melissa] Bring up in [NumPy Community Call](https://www.google.com/url?q=https://hackmd.io/76o-IxCjQX2mOXO_wwkcpg&sa=D&source=calendar&ust=1659794941033190&usg=AOvVaw2nZh-0sHA0ZL2udr6VcS47) later this Wednesday about a having a consistent place for community activity discussion * Comunicate to mailing list, Slack * Point people to that whenever people have quesitons, to existing resources * Should list Youtube channel in more pages, publicize ## Notes for July 18th, 2022 **Present:** Melissa Mendonça (@melissawm), Inessa Pawson (@inessapawson), Sanya Sinha(@ssanya942), Matti Picus (@mattip), Mars Lee (@marsbarlee), Satish Mishra * video from the June 30th NumPy Newcomers’ Hour: https://youtu.be/lHDEr2eCjAM Speaker: Ryan C. Cooper Topic: NumPy in the classroom If you'd like to help with creating timestamps and reviewing/editing/translating subtitles, reach out to Inessa Pawson * Reorganizing the [User guide](https://numpy.org/doc/stable/user/index.html) * Feedback: there are a lot of **outdated pages**; some pages are too detailed for users and should probably be **reorganized**. * Can we come up with an outline for how this should look? (mattip) Do we need to refresh [NEP 44](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html)? * Roadmap: https://numpy.org/neps/roadmap.html#documentation-and-website * Melissa's hypothesis: new users use table of contents so need a straight path to using fundamentals vs seasonsed users use search bar, more like reference in dictionary * User Stories * Made for website, some overlap but should make new user stories for documentation - [x] **Inessa:** make a new tracking issue for making user journies in the documentation. * Can use website analytics? * Only have analytics for numpy.org, not sphinx pages * heat map? * not implemented, may be controversial to implement due to more user tracking, privacy concerns * [plausible tracking](https://plausible.io/): some projects such as pandas have plausible tracking (Google Analytics alteratnive), added via code snippet (thus no overhead) * Example from SciPy: https://github.com/scipy/scipy/blob/main/doc/source/_templates/layout.html * plausible tracking is already implemented on numpy.org * maybe ask NumPy Community meeting and mailing list about plausible tracking * [ ] [name=Melissa] **Make an issue with a top level outline of user guide** * [ ] [name=Melissa] **Make an issue another about user journeys** * [ ] [name=Melissa] **Make a PR about plausible tracking on full docs** * Sanya's experience: uses Google search, or StackOverflow more than documentation. We probably get most traffic via search engine * Inessa: While SEO exists, would not recommend for docs. Instead we can find most searched keywords * Docs are optimized for users, not page clicks * Esnure all pages are searchable, which are enabled by default * In Google Analytics, want to keep a low 'bounce rate', good user stays for 3+ minutes * Inessa: Our bounce rates for numpy.org are good - @inessapawson: documentation sprint with Meekail Zain at SciPy'22: https://hackmd.io/zm65_TjfSpi9DSSkE2o7Iw - **"Zones" of contribution where we identify the missing parts** - Identify missing parts and he is willing to document his learning journey, deliver MVP of what should be visible for people in each zone and expand as new questions come in - Submit an issue? Sharing more widely? Inessa recommend his to join the NumPy meetings - Sharing more widely? As if now, the documentation sprit, Would like more people's opinions about zones - Meekail's opinion: Beginner's is clear but after beginner, much less clear direction - Inessa: State pre-requite knowledge: "This is what you need to know before starting the", ideally in our docs but can link out for now, **example PRs as learning tool** - Matti: For CAPI, specialized knowledge, can be hard because of different backgrounds of contributors - [ ] [name=Matti] **Document some of our disucssion, for example how and why the CAPI maintainers are doing CAPI wor**k - **We need more maintainers** - we are failing to **transition contributors to maintainers** - especially a concern as people get new jobs, other priorities come up - At work setting, all this would be training or shadowing or showing the company codebase - Melissa: Is there any way we can protect mentoring time for the maintainers? - Either part of a grant or hours requirements in the job - Matti: **Livestream PR making**: Here's our Twitch channel, every 2pm on Tuesday, come watch how we do PRs? - Maybe not even need to be planned or annouced, just record on Twitch and release on Twitch - Just record screen, we can add subtitles or voiceover later don't need to entertain crowd - Twitch alternative: **StreamYard**, straight away broadcast on Youtube - [x] [name=Matti] **Talk to Sebestian about streaming the process of working in the code base and PRs** - How much is this documented vs a learning, shadowing process * [name=Rohit] Wording, decisions on `returns` vs `return` * e.g. https://github.com/numpy/numpy/pull/21997#discussion_r922716458 * Personally plugging returns since PEPs don't really apply to us anyway * Comment on the issue if you have opinions * (mattip) commented * [name=Melissa] Updating [the numpy.org design and content improvements tracking issue](https://github.com/numpy/numpy.org/issues/266) * was brought up in roadmap discussion * improvement to installation docs, need to verify * translation * case studies * If you feel like there's no need for some ("we have enough case studies"), then let us know in the issue * [name=Melissa] https://github.com/numpy/numpy/pull/20645 * No one knows? Confusing but no clear resolution * basics indexing, explaining, what triggers advanced indexing, what kinds of objects, indexing arrays (boolean, tuple, etc) * What is the name the object you use in indexing arrays? * "selection object" * "selection tuple" * Would like more people's opinions in the PR * Would only like the word "selection tuple" if a tuple is used, but understandable why if used beyond tuple... * If we don't have a better solution, don't change it * But may remain in limbo for another few years... - [ ] [name=Melissa] **Engage the mailing list for opinions** * Last comment in issue is Ross, seems to know the most on the topic and is a native English speaker * Maybe improve text around the expression instead of changing the expression itself? * Sanya interested in writing, directed to How To for NumPy Tutorials * how do you use this function? * how do you perform common workflows with NumPy? * Sanya can ask NumPy Slack channel for more ## Notes for July 4th, 2022 **Present:** Melissa Mendonça (@melissawm), Angélica Cardozo, Meekail Zain (@micky774), Inessa Pawson (@inessapawson), Sebastian Berg - DOC: Typed NumPY API usage: https://github.com/numpy/numpy/issues/19875#issuecomment-1171221657 Feedback is welcome. - [Quotes ("Mottos") in documentation](https://github.com/numpy/numpy/issues/21858) Feedback is welcome (post directly on GitHub) - DOC: write a how-to guide about using stack/hstack/vstack/concatenate: https://github.com/numpy/numpy/issues/21916 - There are other groups of functions in the documentation that could use how-tos as well. Feel free to open issues requesting this kind of content! - Shout out to Benny and Angelica for working on transcripts and timestamps of the NumPy videos. - Melissa is available to review the transcript in Portuguese. - Having more help with video trascripts would be great. Possible solution: make a call for contributions to the channel subscribers. - [NumPy wiki](https://github.com/numpy/numpy/wiki/) - [Proposal](https://hackmd.io/@melissawm/r1ZhQves5) - Feedback at the meeting: good idea. - DOC: Add a request for explaining the significance of an issue to the issue templates: https://github.com/numpy/numpy/issues/21919 - Good idea - Templates to be modified: a bug report and a feature request. - Inessa will make the changes. Source file: https://github.com/numpy/numpy/tree/main/.github/ISSUE_TEMPLATE - Discussion: how to improve the mid-level/intermediate guidance or documentation for contributors? - Where can people do "meaningful" contributions? - Perhaps video tutorials for how to add a mock method/ufunc/... as a means to introduce audience to machinary surrounding a topic (e.g. how are array methods hooked up, how are modules built/generated, etc.) - Organize documentation with "zones of contribution" (related to contributor journey maps and profiles) - Example zones (just a rough idea): python (internal testing), python (user-facing/API), C (Python C-API), C (ufuncs/backend) - Improve component labels --- ## Notes for June 20, 2022 **Present:** Sebastian Berg, Ryan Cooper, Daniel Wesloh, Mars Lee (@marsbarlee), Inessa Pawson (@inessapawson), Rohit Goswami (@HaoZeke) *Feel free to add items for discussion to this agenda!* * next NumPy Newcomers’ Hour – Thursday, June 30th, 2022 at 4 pm UTC Speaker: Ryan C. Cooper Topic: [NumPy in the classroom](https://hackmd.io/QxWtUwvFTdWZSMZc5SzZbg) <- _WIP meeting materials_ * Issues created for creating timestamps for previous NumPy Newcomer's Hours: * [Analyzing the impact of the lockdown on air quality in Delhi, India](https://github.com/numpy/numpy-newsletter/issues/19) * [Browse and debug NumPy core files with vim, ctags, and cscope](https://github.com/numpy/numpy-newsletter/issues/20) * [Create timestamps of the video Making NumPy accessible with community workshops](https://github.com/numpy/numpy-newsletter/issues/18) * Inessa made these issues in numpy/numpy-newsletter repository as these videos will be higlighted in the upcoming NumPy Newsletter issues. ## Notes for the June 6, 2022 meeting **Present:** Melissa Mendonça (@melissawm), Matti Picus (@mattip), Ryan Cooper (@cooperrc), Mars Lee (@marsbarlee), Rohit Goswami (@haozeke), Sebastian Berg (@seberg) > please add your names (and github handle in brackets). This will make it easier to stay in touch later on issues and pull requests (PRs) 😉 > This is optional since these notes will be recorded in our Github repository. If you'd like you can also paste your answer in the zoom chat 😉 *Feel free to add items for discussion to this agenda!* - [name=Melissa] [Sphinx design migration plans](https://github.com/matplotlib/matplotlib/pull/23169) - Replacement for sphinx-panels - Need less CSS customization- comes with lots of stuff so could clean up own code - No need to act now - [Migration guide](https://sphinx-design.readthedocs.io/en/latest/get_started.html#migrating-from-sphinx-panels) - [name=inessa pawson] NumPy Newcomers’ Hour – June 2nd, 2022 at 4 pm UTC Speaker: Sebastian Berg Topic: Maintaining NumPy: Deprecations Video recording is now available on the NumPy YouTube channel: https://youtu.be/DjiC4-8c5f0 (Remember to subscribe!) - Call for contributions: Create a transcript of the video to be added to the posted YT video and as a separate document for future NumPy documentation (e.g., as a foundation for a how-to guide). - Base off Youtube auto-captions --> `.vtt` and then edit - Hopefully just minor corrections, such as technical terms - Can also add Youtube Timestamps. Helps people find important parts, like headers in documentation - Where to keep transcripts? - Archive in https://github.com/numpy/archive? - [Git LFS](https://git-lfs.github.com/) (Large File Storage)? - Dropbox? - Self-host? --> Resilio Sync has P2P stuff which is free (as long as one of us keeps it on one PC) - Ask other projects if they have unified hosting solution? - Melissa ask NumFOCUS - [name=bhavuk kalra] Should we close these PR's or Keep it around for further discussion - 1) [Revert PR for breathe](https://github.com/numpy/numpy/pull/21270) (Made just in case, If something didn't worked out). 2) [Update Gitpod FAQ](https://github.com/numpy/numpy/pull/21263) - Matti to close this - [name=bhavuk kalra] Mathematical formulae rendering is slow, getting tracked in this issue - [MathJAX rendering is slow](https://github.com/numpy/numpy/issues/21419), suggestions on the test done? - [name=bhavuk kalra] Open for suggestions for a rought draft: [Script for finding and editing the Doc Strings](https://hackmd.io/Uv7TotDjRRKa3iuwf_PpCw) - [name=Melissa] Melissa has written something similar: [How to add examples to NumPy docstrings ](https://hackmd.io/@melissawm/Sy1SGhvD9) - Melissa and Bhavuk to submit a (joint?) PR - [name=Rohit] Which parts of NumPy break SciPy? Should we have a doc explaining that? - e.g. F2PY / maybe Sebastian's changes to dtypes sometimes - Version switcher merged in numpy/doc - [name=Melissa] Any opinions or suggestions on [scipy: Examples missing import numpy as np?](https://github.com/scipy/scipy/issues/13049) - Not directly about us but interesting - "if we have 'import matplotlib as mpl' then why don't we have import numpy as np?" - [name=Melissa] Relevance and role of NumPy wiki - Currently only visible: https://github.com/numpy/numpy/wiki/_history - Wiki mostly used to point at Season of Docs, Summer of Code - Problems - Obscure - Some people not getting notifications - Edit rights and rogue edits - Github edit right options too broad- would like more granular control - Security: spam, marketing, switching links - Solution - Lock it and require wiki changes to be PRs? - Review wiki, see what to keep or move (eg move old meeting notes, archive) and think of new workflow - Open seasonally such as for Season of Docs and let wiki be forum/bulletin board for casual conversation - Melissa will open tracking issue with wiki docs to be moved/deleted --- ## Notes for May 23, 2022 **Present:** Melissa Mendonça, Rohit Goswami (@HaoZeke), Inessa Pawson (@inessapawson), Jake VanderVaate, Ricardo Prins (@ricardoprins), Ross Barnowski, Sebastian Berg, Matti Picus, Brigitta Sipőcz, Lev Maximov, Mukulika Pahari - [name=Rohit] Bumping to the top for some urgency (sorry): - [Mailing list discussion about PDF removal here](https://mail.python.org/archives/list/numpy-discussion@python.org/thread/CCTBM3GONFQWM3DUZRBBT3YYGKXGGPLT/) - Navigation for html is not ideal; this should be fixed irrespective of the PDF generation - There are issues of size of the docker image and other breakages coming from the LaTeX/PDF generation - Rohit makes a point that this is necessary for a possible group of people who are not voicing their needs on github/mailing list - Ross suggests refactoring the generation of the PDF docs to a scheduled CI job that has a designated maintainer (Rohit mentions he could take that up) - Lev mentions that .chm files are also a possibility, as well as Dash (MacOS), Velocity (Windows), Zeal (Linux, Windows), DevDocs (Chrome), Little Drops (Android). - Proposal: come up with a refactoring solution for the next docs meeting --> Rohit will have a demo - [name=Inessa] Translations on crowdin: https://support.crowdin.com/modifying-project-participants-roles/#project-roles - We should check if ownership transfer can be done with Shekhar (Inessa will contact Shekhar) - If not, we can check for an enterprise-level account with NumFOCUS - There was a [PR](https://github.com/numpy/numpy/pull/21522) to add a french readme, should we consider this or maybe consider this after on numpy.org is online? At this time, we are not planning on translating NumPy docs or anything else beyond numpy.org. - [name=bhavuk kalra] Tutorial on Finding and Editing Doc Strings(Rough draft) suggestions :- [Find and Edit Doc Strings](https://hackmd.io/Uv7TotDjRRKa3iuwf_PpCw?view) - [name=Melissa] https://hackmd.io/@melissawm/Sy1SGhvD9 - Melissa and Bhavuk to submit a (joint?) PR - [name=Melissa] https://github.com/numpy/numpy/issues/2067 - [name=Ross] NumPy tutorials CI issues - atari-py/gym is creating problems. The proposal is to freeze the Reinforcement learning notebook and remove those dependencies from requirements/CI. - [name=Brigitta] If the pytest doctesting goes forward with doctestplus, Brigitta would be interested in being involved in the team effort. [SDG proposal for pytest-doctestplus](https://hackmd.io/5uRx7MNhTqSLIFkPnFLY0g) ### For next meeting: - [name=bhavuk kalra] Should we close these PR's or Keep it around for further discussion - 1) [Revert PR for breathe](https://github.com/numpy/numpy/pull/21270) (Made just in case, If something didn't worked out). 2) [Update Gitpod FAQ](https://github.com/numpy/numpy/pull/21263) - [name=bhavuk kalra] Mathematical formulae rendering is slow, getting tracked in this issue - [MathJAX rendering is slow](https://github.com/numpy/numpy/issues/21419), suggestions on the test done? --- ### Useful links - Getting started with building documentation: https://www.numpy.org/devdocs/docs/index.html - How to contribute to the NumPy documentation: https://numpy.org/devdocs/dev/howto-docs.html - Meeting archives: https://github.com/numpy/archive - Mailing list: https://mail.python.org/mailman3/lists/numpy-discussion.python.org/ ## Notes Github Repos: github.com/numpy/numpy - main numpy repo code and docs github.com/numpy/numpy-tutorials - narrative documentation repository github.com/numpy/numpy.org - "front page" of numpy, will be expanded github.com/scipy/scipy github.com/scipy/scipy.org - "front page" of scipy github.com/scipy/docs.scipy.org - one page TOC for documentation https://github.com/scipy/scipy-sphinx-theme - theming for sphinx use - CZI grant for 2020 [goals & deliverables](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167) - CZI grant for 2021 [goals & deliverables](https://figshare.com/articles/online_resource/Improving_usability_and_sustainability_for_NumPy_and_OpenBLAS/13269167)

Syntax	Example	Reference
# Header	Header	基本排版
- Unordered List	Unordered List
1. Ordered List	Ordered List
- [ ] Todo List	Todo List
> Blockquote	Blockquote
Bold font	Bold font
Italics font	Italics font
~~Strikethrough~~	~~Strikethrough~~
19^th^	19^th
H~2~O	H₂O
++Inserted text++	Inserted text
==Marked text==	Marked text
[link text](https:// "title")	Link
![image alt](https:// "title")	Image
`Code`	`Code`	在筆記中貼入程式碼
```javascript var i = 0; ```	`var i = 0;`
:smile:		Emoji list
{%youtube youtube_id %}	Externals
$L^aT_eX$	L^aT_eX
:::info This is a alert area. :::	This is a alert area.