push eventIRkernel/IRkernel

Philipp A

commit sha e42a46fa92a9c05493fa51a0b701daca204d6368

Greatly improve completion

view details

push time in 10 hours

pull request commenttheislab/anndata

[proof of concept/ wip] obsm, varm refactor

Looks great! I just have design considerations (many of which are actually about the current design):

  1. COW, not View. Views propagate changes made to them to the original. Since we have the behavior that we create a copy once we change something, we don’t have views but “copy on write” (aka COW) structures.

  2. Use None instead of 'NoneType'. It’s a bit confusing that None gets used as a synonym for it’s own type, but they decided on that

  3. I don’t understand what idx is for, can you explain? Will it just hold the COW’s subset into the actual structure? Then it should be part of the AlignedCOWMixin, right? You should also should fill in the type parameters for Tuple[?]. Tuple[str, ...] [sic] means that it’s a tuple of strings of unspecified length.

  4. Remove as_dict. dict(thing) will automatically do the same as dict(thing.items()) if thing has an .items() method (which we have, being a MutableMapping)

  5. I don’t think we need to override abstract properties as properties. I think we can just do

    class LayersBase(AlignedMapping):
        attrname = "layers"
    
ivirshup

comment created time in 14 hours

issue openedlandscapeio/requirements-detector

Support pyproject.toml and wheels

pyproject.toml is the new, but only standard way to build packages. Wheels offer a standard way to specify them (using a METADATA file)

So technically the only fool proof way to do this is to build a wheel using pyproject.toml and check that.

But alternatively we can just handle all known users:

if config['build-backend'] == 'flit.api:main':
    flit_meta = config['tool']['flit']['metadata']
    requirements = flit_meta['requires'] + [
        dep
        for extra, deps in flit_meta['requires-extra'].items()
        for dep in deps
    ]
elif config['build-backend'] == 'poetry.masonry.api':
    ... ['tool']['poetry']['dependencies']
...

created time in 15 hours

issue closedIRkernel/IRkernel

jupyter 1.0.0 doesn't provide `jupyter` exectuable

I install IRkernel CRAN release (0.8.15), but IRkernel::installspec() doesn't work.

I find that jupyter 1.0.0 provides jupyter-kernelspec, jupyter-notebook, etc instead of jupyter.

Then I replace all

system2('jupyter', c('kernelspec', ...')

with

system2('jupyter-kernelspec', c(...))

, rebuilt the R package, and it does work.

closed time in 16 hours

dongzhuoer

issue commentIRkernel/IRkernel

jupyter 1.0.0 doesn't provide `jupyter` exectuable

jupyter depends on jupyter_core, which has the script.

It should be installed into ~/.local/bin which you have to add to your $PATH.

dongzhuoer

comment created time in 16 hours

issue commenttidyverse/ggplot2

Get computed statistic from ggplot_build()

Thank you!

flying-sheep

comment created time in 16 hours

issue openedtidyverse/ggplot2

Implement a better way to get the colours or range from a scale

scale_(colour|fill)_gradient, scale_(colour|fill)_gradientn, scale_(colour|fill)_grey, scale_alpha_continuous, and so on all use opaque functions from the scales package instead of some structure that holds their palette/colours.

It would be useful to have a standard way of getting the full input palette/range instead of something like this:

scale_params <- function(scale) {
    proto_env <- environment(scale$palette)
    palette_fun_env <- environment(proto_env$f)
    switch(scale$scale_name,
        gradient =, gradientn = list(colours = palette_fun_env$colours),
        grey = list(colours = scale$palette(2)),  # exec(grey.colors, 2, !!!as.list(palette_fun_env)))
        alpha_c = list(range = palette_fun_env$range),
        ...
    )
}

created time in 16 hours

pull request commentkovidgoyal/calibre

WIP: Port calibre to python 3

I see. And on Arch, we don’t have “foo-common” split packages, so “calibre” would be “calibre-common + calibre-python2” and you’d add “calibre-python3”. Nice! Thank you for your hard work!

flaviut

comment created time in a day

issue commentIRkernel/IRkernel

jupyter 1.0.0 doesn't provide `jupyter` exectuable

What is “jupyter 1.0.0”? a debian package? a pip-installed wheel?

dongzhuoer

comment created time in a day

issue openedtidyverse/ggplot2

Get computed statistic from ggplot_build()

ggplot_build() performs many steps, and the result doesn’t contain the steps between computing and mapping the statistics.

If I do aes(..., alpha = stat(something)), I’d like to be able to get e.g. range(stat(something)).

created time in 2 days

issue commentIRkernel/repr

ERROR while rich displaying an object: Error in if (many_rows) {: argument is of length zero

Thanks for reporting! The newest matter should fix this

ACC839303568

comment created time in 2 days

push eventIRkernel/repr

Philipp A

commit sha 505a052f6b0dd5ccfe57bc88cf7960051065494f

Fix times series display Fixes #117

view details

push time in 2 days

issue closedIRkernel/repr

ERROR while rich displaying an object: Error in if (many_rows) {: argument is of length zero

Environment:

sessionInfo()
R version 3.5.3 (2019-03-11)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 17134)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252    LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                           LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
[1] compiler_3.5.3 tools_3.5.3   
installed.packages()[,c("Package","Version")]
             Package        Version   
base         "base"         "3.5.3"   
base64enc    "base64enc"    "0.1-3"   
boot         "boot"         "1.3-20"  
class        "class"        "7.3-15"  
cluster      "cluster"      "2.0.7-1" 
codetools    "codetools"    "0.2-16"  
compiler     "compiler"     "3.5.3"   
crayon       "crayon"       "1.3.4"   
datasets     "datasets"     "3.5.3"   
digest       "digest"       "0.6.18"  
evaluate     "evaluate"     "0.13"    
foreign      "foreign"      "0.8-71"  
graphics     "graphics"     "3.5.3"   
grDevices    "grDevices"    "3.5.3"   
grid         "grid"         "3.5.3"   
htmltools    "htmltools"    "0.3.6"   
IRdisplay    "IRdisplay"    "0.7.0"   
IRkernel     "IRkernel"     "0.8.15"  
ISLR         "ISLR"         "1.2"     
jsonlite     "jsonlite"     "1.6"     
KernSmooth   "KernSmooth"   "2.23-15" 
lattice      "lattice"      "0.20-38" 
MASS         "MASS"         "7.3-51.1"
Matrix       "Matrix"       "1.2-15"  
methods      "methods"      "3.5.3"   
mgcv         "mgcv"         "1.8-27"  
nlme         "nlme"         "3.1-137" 
nnet         "nnet"         "7.3-12"  
parallel     "parallel"     "3.5.3"   
pbdZMQ       "pbdZMQ"       "0.3-3"   
Rcpp         "Rcpp"         "1.0.1"   
repr         "repr"         "0.19.2"  
rpart        "rpart"        "4.1-13"  
spatial      "spatial"      "7.3-11"  
splines      "splines"      "3.5.3"   
stats        "stats"        "3.5.3"   
stats4       "stats4"       "3.5.3"   
survival     "survival"     "2.43-3"  
tcltk        "tcltk"        "3.5.3"   
tools        "tools"        "3.5.3"   
translations "translations" "3.5.3"   
utils        "utils"        "3.5.3"   
uuid         "uuid"         "0.1-2"   

jupyter notebook --version 5.7.8 (And was installed by Anaconda) Anaconda 1.9.7 Windows 10 Pro

When I tried to execute the r program below on jupyter notebook it showed a mistake on the screen:

a1 <- c(1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20)
b1 <-ts(a1)
b1
ERROR while rich displaying an object: Error in if (many_rows) {: argument is of length zero

Traceback:
1. FUN(X[[i]], ...)
2. tryCatch(withCallingHandlers({
 .     if (!mime %in% names(repr::mime2repr)) 
 .         stop("No repr_* for mimetype ", mime, " in repr::mime2repr")
 .     rpr <- repr::mime2repr[[mime]](obj)
 .     if (is.null(rpr)) 
 .         return(NULL)
 .     prepare_content(is.raw(rpr), rpr)
 . }, error = error_handler), error = outer_handler)
3. tryCatchList(expr, classes, parentenv, handlers)
4. tryCatchOne(expr, names, parentenv, handlers[[1L]])
5. doTryCatch(return(expr), name, parentenv, handler)
6. withCallingHandlers({
 .     if (!mime %in% names(repr::mime2repr)) 
 .         stop("No repr_* for mimetype ", mime, " in repr::mime2repr")
 .     rpr <- repr::mime2repr[[mime]](obj)
 .     if (is.null(rpr)) 
 .         return(NULL)
 .     prepare_content(is.raw(rpr), rpr)
 . }, error = error_handler)
7. repr::mime2repr[[mime]](obj)
8. repr_text.ts(obj)
9. repr_ts_generic(obj, repr_text.matrix, ...)
10. repr_func(m, ..., rows = nrow(m), cols = ncol(m))
11. ellip_limit_arr(obj, ...)
12. arr_partition(a, rows, cols)

However, the program executed well on rstudio ,which showed that:

Time Series:
Start = 1 
End = 20 
Frequency = 1 
 [1]  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20

I tried to reinstall the whole r program and IRkernel by this:

install.packages('IRkernel') 
IRkernel::installspec(user = FALSE)

(See this https://irkernel.github.io/installation/)

Though successfully installed without any warnings and error, it still showed the error above on jupyter notebook.

Still, I tried devtools::install_github('IRkernel/repr') and install.packages(c('IRkernel', 'IRdisplay', 'repr')), but nothing worked for me. (See this https://github.com/IRkernel/repr/issues/100)

Few weeks ago it was nothing wrong with this problem and it did show right, but now it has this kind of exhausting problem……

I know that the jupyter notebook is nothing wrong without this as I executed plot(b1) and it did plot.

I don't know how to solve this problem……

closed time in 2 days

ACC839303568

push eventIRkernel/repr

Philipp A

commit sha a9f07662230a7ebbaff54d151fbb02417c3d930b

fixed matrix-in-df test

view details

push time in 2 days

pull request commenttheislab/scanpy

Minor fixes grab-bag

sounds good from my side

ivirshup

comment created time in 2 days

pull request commentkovidgoyal/calibre

WIP: Port calibre to python 3

I’d be interested, but I don’t see the advantage of depending on the Py2 package. Wouldn’t it be useful to have the pure Python 3 version packaged? Or is the current architecture an a way that the Python 2 version is still used (like unported parts still calling on Python 2 as subprocess or something?)

flaviut

comment created time in 2 days

pull request commentflying-sheep/rust-rst

Filling out the table contents in the document tree.

great, thank you! I know there’s still a lot to do here, so thank you for moving it along!

andreubotella

comment created time in 2 days

issue commenttheislab/scanpy

Contribution of scripts to scanpy main repo

great to hear!

pcm32

comment created time in 2 days

issue commenttheislab/anndata

Investigate lzf compression as default.

https://community.centminmod.com/threads/round-3-compression-comparison-benchmarks-zstd-vs-brotli-vs-pigz-vs-bzip2-vs-xz-etc.17259/

LuckyMD

comment created time in 4 days

issue commenttheislab/scanpy

Contribution of scripts to scanpy main repo

damn, sorry I took so long that you now made all this effort for nothing!

pcm32

comment created time in 7 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

Good! I fixed some typos and improved the examples a bit. Specifically when there’s something to link, they should do it! :class:`pandas.Series` instead of ``pd.Series``

flying-sheep

comment created time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 01b5d78eb00d56b53d607c4030d0f696d2ca5080

Typos, formatting

view details

push time in 7 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

I think you misunderstood a bit.

  1. There’s three types of return sections – prose, tuple, added anndata.obs/var fields.
  2. There’s one style for tuple return sections (like parameters). I implemented an automatic way to handle simple name : type cases of those and manually formatted the rest. I’ll automate the other cases in scanpydoc, then we can remove the manually formatted ones.
  3. There’s 3 styles for anndata.obs/var field return sections. I left them as they are for now, since we have to decide one.

I only reformatted tuple return sections, I neither formatted nor decided on anything about anndata.obs/var field return sections.

I hope I was more clear this time :smile:

flying-sheep

comment created time in 7 days

issue commentpetdance/altbox

Add cw

great!

flying-sheep

comment created time in 7 days

issue commenttheislab/anndata

Conversion to categorical makes None "None"

yikes, definitely. do categoricals support “nan” or some other not-set value?

ivirshup

comment created time in 7 days

push eventflying-sheep/rust-rst

Philipp A

commit sha f00761664cd4e1a23c43914636f8896bb3cf027d

some more simplify

view details

push time in 7 days

issue commentdocutils/docutils

please update the mirror

I also think that this mirror should be one level deeper, i.e. docutils, prest, sandbox, and web should be different projects under the “docutils” organization.

And yes: migration to github or gitlab or so would greatly help community involvement.

return42

comment created time in 7 days

push eventflying-sheep/rust-rst

Philipp A

commit sha df4a83a5639adb21940f1ce43da489ba430621ef

development file

view details

push time in 7 days

issue openedgajus/dindent

Allow using stdin

Your test file_exists('php://stdin') returns FALSE, and file_get_contents doesn’t work with /dev/stdin or -. I suggest that you

  1. get rid of the check (file_get_contents will complain with a better message anyway if it can’t read the file), and
  2. do if ($options['input'] === '-') $options['input'] = 'php://stdin'; for a more natural command line interface

created time in 7 days

delete branch flying-sheep/rust-rst

delete branch : wip-convert

delete time in 7 days

delete branch theislab/scanpy

delete branch : return-formatting

delete time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 0401341877e586e36412db1e178601bf1f545037

Doc overhaul (#610)

view details

push time in 7 days

PR merged theislab/scanpy

Old, simple method for flat return sections

Everything looks like it should:

  • https://scanpy.readthedocs.io/en/return-formatting/api/scanpy.pp.calculate_qc_metrics.html

    Here the rst is weird. Just get rid of the *!

  • https://scanpy.readthedocs.io/en/return-formatting/api/scanpy.pp.normalize_total.html

    Short prose, looks fine.

  • https://scanpy.readthedocs.io/en/return-formatting/api/scanpy.tl.leiden.html

    This should be converted into a definition list as well.

  • https://scanpy.readthedocs.io/en/return-formatting/api/scanpy.tl.dpt.html

    Mixed. Looks great.

No rtype interfering anywhere

+659 -635

12 comments

37 changed files

flying-sheep

pr closed time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 1089b4b76e3e49a34826acd4adb07a3058aa5941

better plot images

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 8f7b98d7e64a84ef400c4ba279881ea85628745d

sphinx 2

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 03e867850db1e25a679d288b6e2be4452ac8c010

improved umap docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 95196668ce6e5056bf3c4e6a2ce89439616fd82e

improved tsne docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 7d293ee77e7e12e82d36326fb88dac5399652d61

improved score_genes_cell_cycle docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha a99d474fb8aeed888f68570309f0fa4a1d975602

prettier phate docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 740761941cdf5d7022b656c638362870c7de3924

improved highly_variable_genes docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha b845c6ef1b8c766e4419a0e3e2adc38547fc8ab6

fixed matplotlib.axes.Axes links

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 46062e5c4d8187510b1ea5f3fb16ca8825be385b

fix filter_rank_genes_groups docs

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha ce6d2b516d76d08541b72796e6869606a562e633

simpler filter_genes doc

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 36528723597852a6e18159b5bc36d73a05a34c75

add colons via CSS

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha f526986ff3c91c3aba369689ba83c1bc4909ce19

whoops 4 – the new order

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 35e8b7aa2e61bd70c37316041c8bb6f066f19df8

whoops³

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 7c212247169219fb69fd6bcfe6dc3ff6c70e28d4

whoops2

view details

push time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha 951a73cf18613fca6546e8d782d1d1b6548772e5

whoops

view details

push time in 7 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

This should work for now. The problem is that this really needs to live in scanpydoc, where we have access to the fancy parsing.

My plan is to merge this now, which has a solution for the simple case of

parameter : some.type
    Description

Later I’ll introduce the same behavior as what scanpydoc does with the parameters:

If the return type is ...) -> Tuple[foo.bar, baz.zab]:, then I’ll check if there’s a section like

one_identifier
    Desc
second_identifier
    Desc

and replace them with the same code as parameters.


This leaves 3 styles:

  1. Prose for a single return value
  2. The above for returning a tuple
  3. “this method adds some anndata fields”

For 3., we have like 3 styles floating around and need to fix one:

  1. **dpt_pseudotime** : :class:`pandas.Series` (`adata.obs`, dtype `float`)
  2. ``adata.obs['louvain']`` (:class:`pandas.Series`, dtype ``category``)
  3. `adata.uns['leiden']['params']` : `dict`
flying-sheep

comment created time in 7 days

push eventtheislab/scanpy

Philipp A

commit sha ccf20f3555a6f240c478c40c1349cc4692665cab

many, many doc fixes

view details

push time in 7 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

OK I have something good. I’ve got an appointment now but will fix the remaining docstrings soon.

flying-sheep

comment created time in 8 days

push eventtheislab/scanpy

Philipp A

commit sha 391cee03d1b812052e8e811144e33ed27f1c75cc

enhanced approach + WIP doc fixes

view details

push time in 8 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

I figured the one-item-thing out: The emitted code is:

:param copy: If an :class:`~anndata.AnnData` is passed, determines whether a copy
             is returned.
:type copy: `bool`, optional (default: `False`)

:returns: AnnData, None
              Depending on `copy` returns or updates `adata` with the corrected data matrix.

And since :returns: is part of a field list, and field lists are defined by the indentation of the block starting in the second line, the additional indentation of the second line is ignored.

So yes, only numpydoc-style sections with one item are affected.

flying-sheep

comment created time in 8 days

Pull request review commenttheislab/scanpy

[WIP] Queries fix

 def biomart_annotations(      Returns     --------    A `pd.DataFrame` containing annotations.+    :class:`pandas.DataFrame`+        Dataframe containing annotations.

sure!

ivirshup

comment created time in 8 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

This is the standard way of writing a numpydoc returns section. […] This solution is dropping support for them.

It certainly shouldn’t, since the definition lists are rendered in other cases. IDK why not here, this should render as a definition list with one item.

However, I don’t like indenting the whole section except for the first line, so in case it always works once there are multiple definition list items, I don’t worry too much here.

Also, do you by chance have another simple solution for having the styling of the return sections similar to the parameters section (what numpydoc did :slightly_smiling_face:)? Bold font and spacings around colons?

I’ll figure it out.

I would remove the , optional statement from the docstrings, as, what we mean with this is "a parameter has a default value". Hence, it's redundant. However, it's consistently used in all of numpy, scipy, sklearn, pandas, etc.

We should definitely put the defaults inline, and I also think the “optional” is redundant. What would it even mean to have “a parameter that isn’t optional but has a default value”?

I’m pretty sure people will understand it.

flying-sheep

comment created time in 8 days

Pull request review commenttheislab/scanpy

[WIP] Queries fix

 def biomart_annotations(      Returns     --------    A `pd.DataFrame` containing annotations.+    :class:`pandas.DataFrame`+        Dataframe containing annotations.

Due to recent doc changes: If you just have one return value, write prose and use a return value annotation.

(same for every other time you changed this in this commit)

ivirshup

comment created time in 8 days

pull request commentsmbache/import

Use name of versioned package in import aliases

No problem!

flying-sheep

comment created time in 8 days

delete branch flying-sheep/import

delete branch : versioned-alias

delete time in 8 days

pull request commentsmbache/import

Use name of versioned package in import aliases

anyone alive?

flying-sheep

comment created time in 8 days

pull request commenttheislab/anndata

[proof of concept/ wip] obsm, varm refactor

Mostly regular instance checks for different things we support:

https://github.com/theislab/anndata/blob/81d10e2f960873c0283b87ae9f7f983100a8d675/anndata/base.py#L51-L61

https://github.com/theislab/anndata/blob/81d10e2f960873c0283b87ae9f7f983100a8d675/anndata/base.py#L831-L838

https://github.com/theislab/anndata/blob/81d10e2f960873c0283b87ae9f7f983100a8d675/anndata/base.py#L743-L758

ivirshup

comment created time in 8 days

pull request commenttheislab/anndata

[proof of concept/ wip] obsm, varm refactor

But all other write operations are meaningless (I think), and that's why DictView, ArrayView etc. exist. They listen to whether you are modifying them and auto-generate a copy if you do so.

FYI: that’s called a “copy on write” structure, or so I think COWDict would be a more fitting name here.

ivirshup

comment created time in 8 days

issue commentcarlin-q-scott/browser-media-players

Links to firefox bugs?

Great, thank you!

flying-sheep

comment created time in 8 days

push eventflying-sheep/phil.red

Philipp A

commit sha f53fa65d57d83c1f8b95ec65eedd9d86e2161ce1

Force LibreOffice LAF

view details

push time in 8 days

issue commenttheislab/scanpy

small bug in sc.pp.filter_cells()

hope I didn't do anything majorly stupid

You did well, good job! You can check the “Files changed” tab of your PR if everything looks like you intended :smile:

LiBuchauer

comment created time in 8 days

pull request commenttheislab/anndata

[proof of concept/ wip] obsm, varm refactor

ABCs are duck typing.

ivirshup

comment created time in 8 days

issue commenttheislab/scanpy

Contribution of scripts to scanpy main repo

Hi @pcm32 I finally got around to this. In the next scanpy version, you should be able to invoke your scripts via e.g scanpy filter-cells, and scanpy -h and scanpy settings will do something too.

pcm32

comment created time in 9 days

delete branch theislab/scanpy

delete branch : cli

delete time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 5a2003ad9d760752dad9afe413528c39005500bb

Implemented a CLI (#604)

view details

push time in 9 days

PR merged theislab/scanpy

CLI

Done!

Fixes #281

+210 -16

0 comment

7 changed files

flying-sheep

pr closed time in 9 days

issue closedtheislab/scanpy

Contribution of scripts to scanpy main repo

Hi there! Thanks for such a great tool.

We are working with a group of people at EBI, Sanger, U. of Freiburg, Earlham (Norwich), the US and Australia in trying to bring different single cell analysis tools to Galaxy and other workflow environments. For tools like Seurat, SC3 or Scanpy, which are normally used as libraries, we are writing an scripting layer for each of these to facilitate their use directly from the shell for well defined functionality (to be called then from whatever workflow environment people want to use). You can get an idea based on what we have here for Seurat for instance https://github.com/ebi-gene-expression-group/r-seurat-scripts. We normally make a conda package out of this called <tool>-scripts. This probably fits well in the initial development phase where we want to be agile in the generation of these, but in the mid and definitely longer term, we would really like to contribute those scripts to the main tool repos, so that they are distributed and installed with them. Would you welcome such contribution (a set of scripts for high level functionality that can be used as executables from a shell) in this git repo?

This is where we are discussing things a bit with this community: https://github.com/galaxyproject/tools-iuc/issues/2057

Current WIP for scanpy scripts (see feature branches): https://github.com/ebi-gene-expression-group/scanpy-scripts/tree/feature/read10x

closed time in 9 days

pcm32

push eventtheislab/scanpy

Philipp A

commit sha dc1c78e3d008b35e8f686aca845799a7c6402f7d

Added entry point

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha cb22003611ac617bbbee9a4f2bc97b1909990772

exclude cli from testing the docs

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 16bb5db928e6897dd523036b13217b292113a3c4

better error if a doc line is missing

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 585cd890ce725c22684e81f1a4380bf6981c2b00

Fix setting numeric verbosity

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 44e7c9e9448f2ef213dd1014b756ce0ec04a7e49

Whoops, we didn’t do that already!

view details

push time in 9 days

push eventtheislab/scanpy

Isaac Virshup

commit sha 21f111984cb421200dfc4f1c88f52fd8d13d8cb9

Use fixtures for downsampling tests

view details

Isaac Virshup

commit sha a4033c6028a175a7f040d446960b5f1bfa2e6d0e

Allow specifying per cell values to downsample to

view details

Isaac Virshup

commit sha 06c36fb98377dcdfd40ba63e47c4f7bdc84221c9

Misc cleanup for downsampling

view details

Alex Wolf

commit sha 873598768a8dc82f016fc2c9eaec469ec1b983d6

Merge pull request #602 from ivirshup/downsample_improvements Downsample improvements

view details

Koncopd

commit sha c7c496232917dcaa45c4512e583604d7056f82e7

rename quantile to fraction in normalize_total, fix docs

view details

Koncopd

commit sha 876286ab27b751fb19ac0d0e639af9f21984f8a9

fix example

view details

Koncopd

commit sha d7e54087b5ba14339759f5033ec6b833db4a6bb7

add test for normalize_total

view details

Koncopd

commit sha f1bdb38863c0370f99bca122a63269572755eabb

add distributed normalize_total test

view details

Alex Wolf

commit sha ea31ff1b63c044ce5ff4c6e3a26c16b7b45174be

Merge pull request #605 from theislab/norml Some changes to normalize_total and tests

view details

Philipp A

commit sha 5b952657b3e88825366df0ea92b96882984b7995

Test new scanpydoc extension See theislab/scanpydoc#7

view details

falexwolf

commit sha 60b3a67eaccafa47eb77b85b83f257b8182fbd28

fixed many docstrings

view details

falexwolf

commit sha 24f34bd909be5fe96b81f8c023d4a08cc6f4d032

updated references

view details

falexwolf

commit sha 52adfdebfe2657f9246227350c4cf53540d02d33

fixed final docstring header

view details

falexwolf

commit sha 5b2c14a0f19ff3f02ca12522780fef692f886b6d

updated some docstrings

view details

falexwolf

commit sha 08111b87ca1274138e6c25963c412b44f7ab57c1

allow prose return sections, remove bullets, allow numpydoc return sections

view details

falexwolf

commit sha 88baa5d0df936c50b6377931314d5fa111e73fbc

docs fixes

view details

falexwolf

commit sha ee498b580e43ed33ebad364d8b2d7be9cb7cc245

updated requirements docs

view details

falexwolf

commit sha d751da6d0ba8a66ed8711cc2aad02513cf1942ef

fixed requirements

view details

Alex Wolf

commit sha 699da5c2c1a52ca52dcb44ed2e42d7aaec906332

Merge pull request #608 from theislab/scanpydoc-flat-return Scanpydoc flat return

view details

falexwolf

commit sha b509978bc84104202255962d523f2e65cface035

updated contributing page

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha e9ab2ba0221f1319d28d0ee3f6fd25ff0378e782

allow other chars in command names

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 8cb52fcd8f5929dcc867e9f4c2f0efa04b73392d

Test in 3.7

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha a0b8313c7e888d67b3f499ed978c9a0798afd382

test improvement

view details

Philipp A

commit sha 5820b59006d25eda350e3378e14374e13787300f

python 3.6

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha b535136edf78dba7a86a26b6c87bb75f4dc543c2

Implemented a CLI

view details

push time in 9 days

pull request commenttheislab/anndata

[proof of concept/ wip] obsm, varm refactor

Good one!

I think we should figure out how to call the type “any matrix type we support”:

>>> import numpy; numpy.ndarray.mro()                                                                                                                                                                                                                  
[numpy.ndarray, object]
>>> import scipy.sparse; scipy.sparse.spmatrix.mro()                                                                                                                                                                                                         
[scipy.sparse.base.spmatrix, object]
>>> import zarr; zarr.Array.mro()
[zarr.core.Array, object]
>>> import xarray; xarray.DataArray.mro()                                                                                                                                                                                                              
[xarray.core.dataarray.DataArray,
 xarray.core.common.AbstractArray,
 xarray.core.common.ImplementsArrayReduce,
 xarray.core.common.DataWithCoords,
 xarray.core.arithmetic.SupportsArithmetic,
 xarray.core.common.AttrAccessMixin,
 object]

So they don’t share any ancestor.

>>> {c: set(dir(c)) & {'toarray', '__array__', 'shape', 'ndim'} for c in [numpy.ndarray, scipy.sparse.spmatrix, zarr.Array, xarray.DataArray]}
{numpy.ndarray: {'__array__', 'ndim', 'shape'},
 scipy.sparse.base.spmatrix: {'ndim', 'shape', 'toarray'},
 zarr.core.Array: {'__array__', 'ndim', 'shape'},
 xarray.core.dataarray.DataArray: {'__array__', 'ndim', 'shape'}}

but they do all have shape, ndim, and either __array__ or toarray!

  • So we could accept anything that has that.
  • When slicing the whole thing and that fails, we need to emit a good error message telling people which obs exact entry in which annotation failed to be slicable.
ivirshup

comment created time in 9 days

pull request commentDefinitelyTyped/DefinitelyTyped

[copy-webpack-plugin] fix transform parameter

True. Welp, a new PR?

dsifford

comment created time in 9 days

issue commentpetdance/altbox

Start on non-alternative tools

See also #12.

I think rq needs to be there too.

petdance

comment created time in 9 days

issue openedpetdance/altbox

xsv

created time in 9 days

issue openedpetdance/altbox

Add cw

cw is a faster wc written in Rust. It supports counting words, lines, (utf8-)characters and bytes

created time in 9 days

issue openedpetdance/altbox

Add lsd

Like exa, it’s a ls replacement written in Rust, with different strenghts and weaknesses.

It’s a bit closer to ls’ options, and displays icons (using font-awesome/nerdfont codepoints).

https://github.com/Peltoche/lsd

created time in 9 days

pull request commentjupyter/nbconvert

Mention formats in --to documentation

great!

flying-sheep

comment created time in 9 days

delete branch flying-sheep/nbconvert

delete branch : patch-1

delete time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 7c450114ff376d2c6095f03855927cd1428a5f17

Use updated version

view details

push time in 9 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

Ah, the problem was that the string actually contained that return type!

Theoretically that should have parsed as a 1-element definition list, but I don’t think it’s a problem that it doesn’t.

flying-sheep

comment created time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha fae1c29c6d0fe2007aaf3a898dc205d27336d3b9

Fix calculate_qc_metrics docstring

view details

push time in 9 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

Well, it wasn’t that one anyway. I’ll figure it out

flying-sheep

comment created time in 9 days

pull request commenttheislab/scanpy

Old, simple method for flat return sections

Wait, aren’t the docs supposed to be like this?

def myfunc():
    """\
    First line
    ...
    """
flying-sheep

comment created time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha fa141ac1a7369bb29a11a942e5af6dee0bfd32bc

Fix doc indentation

view details

push time in 9 days

push eventtheislab/scanpy

Philipp A

commit sha 48f95da82ac8694a94f80e48d12f676be1e30331

grammar

view details

push time in 9 days

push eventtheislab/scanpy

Alex Wolf

commit sha c5e3b17440f84dc6cf874cb3680c481e370191ab

updated contributing docs section

view details

Alex Wolf

commit sha dceb861873d1a2909c325fa830ee22aa4a01d054

fixed dendrogram docstring

view details

Philipp A

commit sha 108121cb915e39658643d70494ef7595a397172a

Merge branch 'master' into return-formatting

view details

push time in 9 days

PR opened theislab/scanpy

Old, simple method for flat return sections
+30 -76

0 comment

2 changed files

pr created time in 9 days

delete branch theislab/scanpy

delete branch : scanpydoc-flat-return

delete time in 9 days

create barnchtheislab/scanpy

branch : return-formatting

created branch time in 9 days

pull request commenttheislab/anndata

use obs and var when creating adata from df

Great, thank you very much!

fabianrost84

comment created time in 10 days

more