Revert "Revert "Use HighLevelGraph layers everywhere in collections (…#6707
Merged
jrbourbeau merged 1 commit intodask:masterfrom Oct 8, 2020
Merged
Revert "Revert "Use HighLevelGraph layers everywhere in collections (…#6707jrbourbeau merged 1 commit intodask:masterfrom
jrbourbeau merged 1 commit intodask:masterfrom
Conversation
Member
|
Let's wait until we do a release. There were some issues last time around
with string keys in subgraphcallables. I think that we might want to do a
small mid-week release for this as well as anything else that feels
critical, and then start merging in the HLG fixes.
…On Tue, Oct 6, 2020 at 4:35 AM Tom Augspurger ***@***.***> wrote:
…#6510 <#6510>)" (#6697
<#6697>)"
This reverts commit e09d8d9
<e09d8d9>
.
cc @madsbk <https://github.com/madsbk>
------------------------------
You can view, comment on, or merge this pull request online at:
#6707
Commit Summary
- Revert "Revert "Use HighLevelGraph layers everywhere in collections
(#6510)" (#6697)"
File Changes
- *M* dask/array/optimization.py
<https://github.com/dask/dask/pull/6707/files#diff-10351144025e8c061b086321d0d9c9b6>
(33)
- *M* dask/blockwise.py
<https://github.com/dask/dask/pull/6707/files#diff-99ed6f84a26a1697121f5e38e8b555c1>
(55)
- *M* dask/core.py
<https://github.com/dask/dask/pull/6707/files#diff-1ede41868431981d4249ebdea17d91b4>
(28)
- *M* dask/dataframe/io/parquet/core.py
<https://github.com/dask/dask/pull/6707/files#diff-c9c912457d8bd444a48523c4b8d1d58e>
(34)
- *M* dask/dataframe/optimize.py
<https://github.com/dask/dask/pull/6707/files#diff-6737b31c383e68893904fdd3c986b6c7>
(27)
- *M* dask/highlevelgraph.py
<https://github.com/dask/dask/pull/6707/files#diff-813eb8bcdb581604b1788c0e38305049>
(229)
- *M* dask/tests/test_highgraph.py
<https://github.com/dask/dask/pull/6707/files#diff-8cab368f81ee56fe90ad3f18200dea56>
(19)
Patch Links:
- https://github.com/dask/dask/pull/6707.patch
- https://github.com/dask/dask/pull/6707.diff
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#6707>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AACKZTEMPOHEII4W3KBHXZ3SJL6JTANCNFSM4SF4ZNKA>
.
|
jrbourbeau
reviewed
Oct 7, 2020
Member
jrbourbeau
left a comment
There was a problem hiding this comment.
Playing around with this PR locally it looks like we can still produce HighLevelGraphs that have raw dictionaries in their layers which causes HighLevelGraph.get_dependencies to fail. For example:
import dask.array as da
x = da.ones(10, chunks=(2,))
dsk = x.__dask_graph__()
# Print out HighLevelGraph layers for x
from pprint import pprint
pprint(dsk.layers)
# Get dependencies for all the keys
deps = dsk.get_dependencies()
print(f"{deps = }")raises and AttributeError:
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-1-61e7febb21f5> in <module>
8
9 # Get dependencies for all the keys
---> 10 deps = dsk.get_dependencies()
11 print(f"{deps = }")
~/projects/dask/dask/dask/highlevelgraph.py in get_dependencies(self)
369 self.key_dependencies = {}
370 for layer in self.layers.values():
--> 371 self.key_dependencies.update(layer.get_dependencies(all_keys))
372
373 return self.key_dependencies
AttributeError: 'dict' object has no attribute 'get_dependencies'
I see there's a HighLevelGraph._fix_hlg_layers_inplace method which ensures all layers are of type Layer. Should we call that at the end of HighLevelGraph.__init__ to make sure we always have a uniform layer type? @madsbk I'd be curious to hear what you think -- I also could be totally missing something
Contributor
Good point, I totally agree :) I think we should:
Afterwards, we can look at merging:
After which, we can begin the working of sending HLGs directly to the scheduler in Distributed:
|
jrbourbeau
approved these changes
Oct 8, 2020
6 tasks
Member
|
Thanks all for coordinating! |
2 tasks
kumarprabhu1988
pushed a commit
to kumarprabhu1988/dask
that referenced
this pull request
Oct 29, 2020
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
…#6510)" (#6697)"
This reverts commit e09d8d9.
cc @madsbk