Re: Beancount on PyPi
> I don't think I will wait to close all the tickets before I make numbered > releases. > It's been pretty stable and I don't have enough bandwidth these days. > > Here's what I see is next in terms of "small stuff" that would be part of > near upcoming 2.x releases: > This is small-ish stuff that gets on my own nerves regularly. I think that laying out the "roadmap" as clearly as you did below does help contributors to pick things that they then know that you are also interested in, and work on those things. Personally I think now that the project feels like gaining more and more followers/users the amount of PRs, issues, etc. likely won't decline. Automatically converting the docs to HTML format is a very good step for improving discoverability (HTML can be search-indexed, etc.), as well as keeping to Google Docs as the "UI"/mechanism for editing them keeps the bar low for contributions to the docs. As for the code, moving to git (and github.com/beancount/beancount? We already "own" that org) would further lower the barriers to help/contribute for many more people. Please (you, dear reader) do not take this as a starting-point for another lengthy discussion about this. I just want to friendly remind you, Martin, what possible ways of lowering the amount of work beancount takes up could be. - Dominik > Am 24.03.2018 um 06:30 schrieb Martin Blais : > > How would you feel about me baking an official 2.0 stamp on the current tip? > I was waiting to finish the doc integration to do this and adopt semantic > versioning. > > I don't think I will wait to close all the tickets before I make numbered > releases. > It's been pretty stable and I don't have enough bandwidth these days. > > Here's what I see is next in terms of "small stuff" that would be part of > near upcoming 2.x releases: > - Integrate Dominik's docs as a static copy. > - Adding coherent cmdline options to disable the caches (pickle cache, price > source cache, implement import cache for costly conversions). > - Remove the Holding class (less code); replace all its functionality by > aggregation of positions. Rewrite all the reports with SQL queries only. > - Promote list_holdings.py (the script I use myself) to the main codebase and > share an example Google Sheets doc to track assets (that's what I've been > doing for years, works great). > - Figure out what to do about speed (we need something like a 3x speedup IMO). > - Do something about price sources now that G and Y finance are gone (perhaps > just provide a helpers library instead of a tool, not sure) > This is small-ish stuff that gets on my own nerves regularly. > > Here's some stuff that's being requested regularly by others that I also > would like to work on: > - Adding options to override / take priority over inference for precision and > allow you to set the num digits to render directly. > - Take a solid look at the CSV importer contributions waiting in PRs. > - Implement AVERAGE booking. > - Implement query assertions to have a good budget story / take time to check > out Fava's custom directive for that, maybe integrate. > - Supporting UTF8 > and of course, slowly work my way through all those tickets and pending PRs. > > There's work for three dozen lifetimes (I don't suggest wading through the > TODO file, you'll need courage and it's organized like a 14 year old's > bedroom). Overall I'd like to reduce the amount of code (deprecate Holdings, > perhaps delete bean-web (because Fava, if it can be baked), factor out SQL > query client to a configurable row-based query tool that works on any data > source) and focus on doing more with a smaller body of code (e.g. compute > returns, support complex risk calculations (I hold options), implement pivot > table in SQL for e.g. monthly reports, make the shell so powerful it's the > only way to generate reports, simplify the tags/links schema, etc.). > > > > > On Fri, Mar 23, 2018 at 10:00 AM, Martin Michlmayr wrote: > 2.0rc1 is almost a year old. Would it be possible to cut an rc2 since > there have been a lot of changes in the meantime? > > This would help those not installing from mercurial directly. > > Thanks, > > Reference: > https://pypi.org/project/beancount/#history > -- > Martin Michlmayr > http://www.cyrius.com/ > > -- > You received this message because you are subscribed to the Google Groups > "Beancount" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to beancount+unsubscr...@googlegroups.com. > To post to this group, send email to beancount@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/beancount/20180323140005.4y4rquuqigi7mgts%40jirafa.cyrius.com. > For more options, visit https://groups.google.com/d/optout. > > > -- > You received this message because you are subscribed to the Google Groups > "Beancount" group. > To unsubscribe from this group and stop receiving e
Re: Documentation is now read-only
This looks great! Thanks for working on this! :) > It's incomplete, but you can see how pretty darn close it is (it's readable, > but flawed). Except for the code blocks (they should be in a monospaced font, as otherwise it would be confusing for the reader), and the sidebar (can be configured in Sphinx, afaik), this looks as pretty as I hoped for. How much work do you plan to put in until you "release" this? (As it looks like not much is needed for "release-readyness", imho.) > Am 24.03.2018 um 05:35 schrieb Martin Blais : > > Thx Dominik. > > BTW, here's what the current automated conversion results looks like in your > static skeleton: > http://furius.ca/tmp/beancount-docs/ > > This was entirely produced by the code I wrote here, no manual changes: > https://bitbucket.org/blais/beancount/src/tip/experiments/docs_rst/ > > It's incomplete, but you can see how pretty darn close it is (it's readable, > but flawed). > > Quick comments: > - I haven't copied all the docs, just the ones you had linked in, it would be > easy to include more in the structure > - Some - but not all - of the code blocks are erroneous, they have an extra > duplicated unindented last line (probably a bug in convert_docs.py) > - Some of the **bold** and *italic* strings don't render as such, I don't > know why (input looks right) > - Some of the list items somehow have their indentation badly converted by > Pandoc > - The index on the left shows sections for the entire set of documents - this > is too much, should show sections for each document only > > I think post-processing the rst from Pandoc it would be possible to fix all > these things to make it a fully automated conversion. > > > > > On Thu, Mar 15, 2018 at 2:19 AM, Dominik Aumayr wrote: >> - Eventually I'd like to merge that into Beancount itself and maintain it >> (with Dominik's permission). > > Permission hereby granted :-) > > Am 14.03.2018 um 06:08 schrieb Martin Blais : > >> I just cleaned up the hacky conversion code a bit. >> >> If anybody would like to have a go at completing this, the source is here: >> https://bitbucket.org/blais/beancount/src/tip/experiments/docs_rst/?at=default >> >> Instead of running the download_docs.py script, which might require a Google >> API token, I made an archive of what it downloads here: >> http://furius.ca/tmp/beancount-docs-downloaded/docs.tar.gz l (28MB) >> >> See the Makefile. >> What remains to be done is >> - Check that the output converts nicely >> - If not, find ways to automatedly fix it during the conversion, I'm not >> 100% sure that's possible (I'm hoping) >> - Integrate it in Dominik Aumayr's static docs and look at the converted text >> - Eventually I'd like to merge that into Beancount itself and maintain it >> (with Dominik's permission). >> >> >> >> On Sun, Mar 11, 2018 at 12:56 PM, Martin Blais wrote: >> Not planning to move to rst as an input format. >> I still would use GDocs as input; just request comment access and I'll >> accept liberally. >> The idea would be to fully automate the conversion. >> (The only reason I made the docs RO was to avoid the numerous and frequent >> fingerfarts.) >> >> On Sat, Mar 10, 2018 at 6:43 PM, Martin Michlmayr wrote: >> * Martin Blais [2018-02-20 00:07]: >> > Anyhow, I'll try to finish this and convert all the docs to rst at some >> > point. >> >> Once the rst documents exist, are you going to accept pull requests >> with changes against the rst (and apply them to your canonical source) >> or how should people interested in working on docs submit changes? >> >> -- >> Martin Michlmayr >> http://www.cyrius.com/ >> >> >> >> -- >> You received this message because you are subscribed to the Google Groups >> "Beancount" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to beancount+unsubscr...@googlegroups.com. >> To post to this group, send email to beancount@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/beancount/CAK21%2BhNTudNTXeuBaATdfFCk0q2_wV3HJd%3DZ8ZV5K3cgVm3Org%40mail.gmail.com. >> For more options, visit https://groups.google.com/d/optout. > > -- > You received this message because you are subscribed to the Google Groups > "Beancount" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to beancount+unsubscr...@googlegroups.com. > To post to this group, send email to beancount@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/beancount/9CBA7B3A-292A-4803-BD76-C8C4B1AA90B2%40aumayr.name. > > For more options, visit https://groups.google.com/d/optout. > > > -- > You received this message because you are subscribed to the Google Groups > "Beancount" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to beancount+unsubscr...@googlegroups.com. > To post to this group, send email to beancount@go
Re: Beancount on PyPi
On Sat, Mar 24, 2018 at 01:30:09AM -0400, Martin Blais wrote: > How would you feel about me baking an official 2.0 stamp on the > current tip? I was waiting to finish the doc integration to do this > and adopt semantic versioning. > > I don't think I will wait to close all the tickets before I make > numbered releases. It's been pretty stable and I don't have enough > bandwidth these days. The current version feels indeed stable enough to be stamped 2.0. But. Pushing a new release to PyPi will probably make the tip version (as opposed to the last rc1) get some more exposure from users that are not using tip. So it might make sense to release current tip as rc2 to PyPi, wait 2 weeks (to see if anything important shows up), then stamp 2.0. My (semantically versioned) 0.0.2 EUR, Cheers -- Stefano Zacchiroli . z...@upsilon.cc . upsilon.cc/zack . . o . . . o . o Computer Science Professor . CTO Software Heritage . . . . . o . . . o o Former Debian Project Leader & OSI Board Director . . . o o o . . . o . « the first rule of tautology club is the first rule of tautology club » -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20180324104631.GD25046%40upsilon.cc. For more options, visit https://groups.google.com/d/optout.
Re: moving to GitHub?
* Dominik Aumayr [2018-03-24 11:10]: > As for the code, moving to git (and github.com/beancount/beancount? > We already "own" that org) would further lower the barriers to > help/contribute for many more people. Please (you, dear reader) do > not take this as a starting-point for another lengthy discussion > about this. I just want to friendly remind you, Martin, what > possible ways of lowering the amount of work beancount takes up > could be. I was actually going to ask Martin about this in private (in private because I didn't want to start a bike-shedding discussion). I've submitted some minor pull requests and while hg is not too hard if you know git, remembering the right hg commands and signing up for bitbucket is a small barrier that imho could easily be avoided by moving to GitHub. If Martin says no, we should accept that without further discussion. But if it's a maybe or yes, I'd be happy to look into the migration (unless someone has experience with this already). There is a script that can be used to import bitbucket issues into GitHub: https://github.com/jeffwidman/bitbucket-issue-migration/blob/master/migrate.py And the mercurial repo is mirrored already. -- Martin Michlmayr http://www.cyrius.com/ -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20180324151416.vmhnrm6j3w6yyy2c%40jirafa.cyrius.com. For more options, visit https://groups.google.com/d/optout.
Re: Beancount on PyPi
* Martin Blais [2018-03-24 01:30]: > How would you feel about me baking an official 2.0 stamp on the current tip? I was actually planning to send an email to ask about the status of 2.0, what needs to be done and how people can help. :) > I don't think I will wait to close all the tickets before I make numbered > releases. > It's been pretty stable and I don't have enough bandwidth these days. > > Here's what I see is next in terms of "small stuff" that would be part of > near upcoming 2.x releases: Personally, I'd make another rc release now, prioritize some items for 2.0 (as you did), encourage people to submit fixes for 2.0 and make 2.0 in a few months. This assumes that you can make some time to review patches, accept pull requests and possibly fix some of the harder issues yourself. Of course, you could release 2.0 now and do a 2.x in a few months but I feel some stuff is missing from what I'd call a "2.0". I'm still exploring beancount but for example UTF-8 support in account names is crucial for some users (like our French friends :). So maybe do an rc now and set a date for the 2.0 release (3 months from now?) to give people a chance to get fixes/features in without delaying 2.0 for too long. > - Do something about price sources now that G and Y finance are gone > (perhaps just provide a helpers library instead of a tool, not sure) Someone on this list posted the following in December https://github.com/hoostus/beancount-price-sources which supports Morningstar and OpenExchange. (I also saw anotehr OpenExchange implementation on GitHub.) There's a pull request for Quandl. I haven't used an price source yet but maybe these would be alternatives. Thanks for creating beancount, btw! -- Martin Michlmayr http://www.cyrius.com/ -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20180324152726.war4cucvzomxn7zr%40jirafa.cyrius.com. For more options, visit https://groups.google.com/d/optout.
Re: Documentation is now read-only
On Sat, Mar 24, 2018 at 12:35:23AM -0400, Martin Blais wrote: > BTW, here's what the current automated conversion results looks like in > your static skeleton: > http://furius.ca/tmp/beancount-docs/ This is just ... wow ... amazing work Martin, thanks for doing this! My only comment (in the same spirit of Dominik's point about lowering the barrier for contributions) is to document in the doc itself how to contribute. I understand you want comments/edit suggestion via Google Docs. And each sphinx document already points to the corresponding Google Docs origin. I suggest prefixing each link with something like "to suggest edits to this document go to ...". And I'm assuming you will re-enable suggestions on the Google Docs once the sphinx documents will be advertised as the main entry points for users, right? Cheers -- Stefano Zacchiroli . z...@upsilon.cc . upsilon.cc/zack . . o . . . o . o Computer Science Professor . CTO Software Heritage . . . . . o . . . o o Former Debian Project Leader & OSI Board Director . . . o o o . . . o . « the first rule of tautology club is the first rule of tautology club » -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20180324163324.GD31031%40upsilon.cc. For more options, visit https://groups.google.com/d/optout.
Re: Beancount on PyPi
2018-03-24 06:46:31 Stefano Zacchiroli : > On Sat, Mar 24, 2018 at 01:30:09AM -0400, Martin Blais wrote: > > How would you feel about me baking an official 2.0 stamp on the > > current tip? I was waiting to finish the doc integration to do this > > and adopt semantic versioning. > > > > I don't think I will wait to close all the tickets before I make > > numbered releases. It's been pretty stable and I don't have enough > > bandwidth these days. > > The current version feels indeed stable enough to be stamped 2.0. But. > Pushing a new release to PyPi will probably make the tip version (as > opposed to the last rc1) get some more exposure from users that are not > using tip. So it might make sense to release current tip as rc2 to PyPi, > wait 2 weeks (to see if anything important shows up), then stamp 2.0. > > I agree with Stefano. We have so many changes since RC1, we should stamp another "release candidate" before 2.0. If Martin is okay with moving the repository to GitHub, as Dominik and many others suggest, we could release RC2 directly to GitHub. This could greatly speed up the progress towards 2.0 because of more contributors. -- Zhuoyun Wei -- You received this message because you are subscribed to the Google Groups "Beancount" group. To unsubscribe from this group and stop receiving emails from it, send an email to beancount+unsubscr...@googlegroups.com. To post to this group, send email to beancount@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/beancount/20180325031409.GA9669%40tarball.wzyboy.org. For more options, visit https://groups.google.com/d/optout. signature.asc Description: PGP signature