hail-is/hail-dev

Konrad Karczewski

@konradjk

is min_partitions in import_table not working?

Laurent Francioli

@lfrancioli

that could explain it :)

Daniel King

@danking

@gtiao it looks like you might be loading all the data into one partition?

gtiao

@gtiao

I thought that was what I was trying to avoid by

ht = hl.import_table('gs://gnomad/variant_qc/temp/friedman_cnn_scores.tsv.gz', force_bgz=True, min_partitions=1000, impute=True)

Daniel King

@danking

min_block_size controls how small input file blocks can be

I also think that would avoid it.

gtiao

@gtiao

Maybe force_bgz doesn’t work with min_partitions?

It is actually bgzipped

Daniel King

@danking

lemme spin up a little cluster to poke at that file

Daniel King

@danking

blah. we need to make 0.2 the default such a long turn around time if I forget

Daniel King

@danking

so its definitely being loaded in one partition

which is obviously bad and wrong, and I’ll try to figure out why

gtiao

@gtiao

OK, cool — thanks for looking into it!

Daniel King

@danking

@gtiao you’re on latest master right?

Daniel King

@danking

@gtiao yeah it’s force_bgz being broken somehow, if you can rename the file that will bypass the issue for now

gtiao

@gtiao

Great — I will do that. I’ve been using a Konrad jar (gs://konradk/jars/hail-6d4d50458.jar) but I don’t recall what the specific issue was that we were trying to address with that

Konrad Karczewski

@konradjk

am i to presume that if i see this in my file:

 'gs://gnomad/annotations/hail-0.2/ht/exomes/gnomad.exomes.family_stats.ht/rows/parts/part-04249-15-4249-0-0e67bae1-c1d2-5e25-aad1-eb8c419bbdbe',
 'gs://gnomad/annotations/hail-0.2/ht/exomes/gnomad.exomes.family_stats.ht/rows/parts/part-04249-15-4249-1-8e47c4bd-6382-0449-a9c9-7f3d29ea1511',

that the later one is the correct one?

klaricch

@klaricch

any thoughts how to follow up on the error below? ld_prune had worked on an exome matrix table but then I joined it with data from an array matrix table and lost a lot of variants and kept only GT as an entry field. not sure if i need to skip that join.
mm_test = hl.ld_prune(mm.GT,r2=0.1)

FatalError: ArrayIndexOutOfBoundsException: 6

Java stack trace:
java.lang.ArrayIndexOutOfBoundsException: 6
    at is.hail.methods.LocalLDPrune$.apply(LocalLDPrune.scala:294)
    at is.hail.methods.LocalLDPrune.apply(LocalLDPrune.scala)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
    at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
    at py4j.Gateway.invoke(Gateway.java:280)
    at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
    at py4j.commands.CallCommand.execute(CallCommand.java:79)
    at py4j.GatewayConnection.run(GatewayConnection.java:214)
    at java.lang.Thread.run(Thread.java:748)

Hail version: devel-15eaf7588401
Error summary: ArrayIndexOutOfBoundsException: 6

cseed

@cseed

@klaricch It looks like a bug in our end. Can you open an issue on the github repo? Thanks!

It looks like it should be straightforward to fix.

maccum

@maccum

hail-is/hail#3735 should fix that bug once it goes in @klaricch

klaricch

@klaricch

ok thanks!

Laurent Francioli

@lfrancioli

I'm having troubles with AssertionError on Tables:

ht.describe()
----------------------------------------
Global fields:
    None
----------------------------------------
Row fields:
    'v1_idx': int32 
    'v2_idx': int32 
----------------------------------------
Key: ['v1_idx', 'v2_idx']
----------------------------------------
ht.show()
Traceback (most recent call last):
  File "/anaconda3/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2910, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-53-73b5a6c78295>", line 1, in <module>
    ht.show()
  File "/Users/laurent/tools/hail-release/devel/hail.zip/hail/typecheck/check.py", line 547, in wrapper
    return f(*args_, **kwargs_)
  File "/Users/laurent/tools/hail-release/devel/hail.zip/hail/table.py", line 1169, in show
    print(self._show(n,width, truncate, types))
  File "/Users/laurent/tools/hail-release/devel/hail.zip/hail/table.py", line 1172, in _show
    return self._jt.showString(n, joption(truncate), types, width)
  File "/Users/laurent/tools/spark-2.2.0-bin-hadoop2.7/python/lib/py4j-0.10.4-src.zip/py4j/java_gateway.py", line 1133, in __call__
    answer, self.gateway_client, self.target_id, self.name)
  File "/Users/laurent/tools/hail-release/devel/hail.zip/hail/utils/java.py", line 196, in deco
    'Error summary: %s' % (deepest, full, hail.__version__, deepest)) from None
hail.utils.java.FatalError: AssertionError: assertion failed
Java stack trace:
java.lang.AssertionError: assertion failed
    at scala.Predef$.assert(Predef.scala:156)
    at is.hail.expr.ir.TypeCheck$.apply(TypeCheck.scala:78)
    at is.hail.expr.ir.TypeCheck$.apply(TypeCheck.scala:7)
    at is.hail.expr.ir.Emit$.emit(Emit.scala:42)
    at is.hail.expr.ir.Emit$.apply(Emit.scala:28)
    at is.hail.expr.ir.Compile$.apply(Compile.scala:49)
    at is.hail.expr.ir.Compile$.apply(Compile.scala:31)
    at is.hail.expr.ir.Compile$.apply(Compile.scala:62)
    at is.hail.expr.TableExplode.execute(Relational.scala:2201)
    at is.hail.expr.TableUnkey.execute(Relational.scala:1883)
    at is.hail.expr.TableMapRows.execute(Relational.scala:2090)
    at is.hail.expr.TableKeyBy.execute(Relational.scala:1846)
    at is.hail.expr.TableMapRows.execute(Relational.scala:2090)
    at is.hail.table.Table.value$lzycompute(Table.scala:243)
    at is.hail.table.Table.value(Table.scala:238)
    at is.hail.table.Table.x$5$lzycompute(Table.scala:246)
    at is.hail.table.Table.x$5(Table.scala:246)
    at is.hail.table.Table.rvd$lzycompute(Table.scala:246)
    at is.hail.table.Table.rvd(Table.scala:246)
    at is.hail.table.Table.take(Table.scala:961)
    at is.hail.table.Table.showString(Table.scala:1002)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at py4j.reflection.MethodInvoker.invoke(MethodInvoker.java:244)
    at py4j.reflection.ReflectionEngine.invoke(ReflectionEngine.java:357)
    at py4j.Gateway.invoke(Gateway.java:280)
    at py4j.commands.AbstractCommand.invokeMethod(AbstractCommand.java:132)
    at py4j.commands.CallCommand.execute(CallCommand.java:79)
    at py4j.GatewayConnection.run(GatewayConnection.java:214)
    at java.lang.Thread.run(Thread.java:745)
Hail version: devel-10a75bb57a6f
Error summary: AssertionError: assertion failed

clues?

gtiao

@gtiao

If I’ve left out a header on an imported file, can I add colnames by doing table_result = table1.rename({'C1' : ‘newcolname1', 'C2' : ‘newcolname2’})?

maccum

@maccum

@lfrancioli that’s a bug. can you make an issue in github?

Laurent Francioli

@lfrancioli

:thumbsup:

Konrad Karczewski

@konradjk

@lfrancioli what was your pipeline?

Laurent Francioli

@lfrancioli

    def variant_pairs_ht(mt, row_groups):
        mt = mt.add_row_index(name='_row_idx')
        mt = mt.add_col_index(name='_col_idx')

        mt = mt.group_rows_by(*row_groups).aggregate(
            variant_pairs_entry=hl.set(
                hl.agg.collect(
                    hl.tuple([mt._col_idx, mt._row_idx])
                )
                    # [col_idx, [row_idx]]
                    .group_by(lambda x: x[0])
                    # [[col_idx, row_idx), (col_idx, row_idx), ...],  ...]
                    .values()
                    # [[row_idx, row_idx, ...],  ...]
                    .map(lambda x: x.map(lambda y: y[0]))
                    .flatmap(lambda x: hl.range(0, hl.len(x))
                               .flatmap(lambda i1: hl.range(i1 + 1, hl.len(x))
                                        .map(lambda i2: hl.tuple([x[i1], x[i2]]))))
            )
        )

        mt.describe()

        ht = mt.annotate_rows(
            variant_pairs=hl.agg.take(mt.variant_pairs_entry, 1)[0]
        ).rows()

        ht = ht.explode('variant_pairs')
        ht = ht.key_by(v1_idx=ht.variant_pairs[0], v2_idx=ht.variant_pairs[1])
        return ht.select()

Konrad Karczewski

@konradjk

good lord

gtiao

@gtiao

I just got a warning on a pipeline:

UserWarning: The mt[<row keys>, :] syntax is deprecated, and will be removed before 0.2 release.
  Use one of the following instead:
    mt.rows()[<row keys>]
    mt.index_rows(<row keys>)
  ht = ht.annotate(info=hl.struct(AC=mt[ht.key, :].info.AC, a_index=mt[ht.key, :].a_index))

Does this mean the new syntax should read:

ht = ht.annotate(info=hl.struct(AC=mt.rows()[ht.key].info.AC, a_index=mt.rows()[ht.key].a_index))

Konrad Karczewski

@konradjk

yes, but don't do that yet

gtiao

@gtiao

Why not?

Konrad Karczewski

@konradjk

table joins are still not quite as fast as MT

gtiao

@gtiao

So keep till it breaks?

Konrad Karczewski

@konradjk

yes

gtiao

@gtiao

Is there a way to join two VCFs in Hail 0.2?

outer join

basically, merge

Konrad Karczewski

@konradjk

outer join on what?

samples, variants, or both?

gtiao

@gtiao

Both

Konrad Karczewski

@konradjk

i don't think so

outer on one, inner on the other is possible with union_cols and union_rows

gtiao

@gtiao

Is there a technical reason why you can’t do outer on both?

“you” in the general sense, of course

jkeebler

@jkeebler

outer join requires some knowledge of sample non-carrier status (missing vs homozygous ref) for variants seen in one VDS but not the other .. not possible w/out going back to BAMs, incorporating the TileDB storage, or supporting gVCF .. not sure where those are on the Hail roadmap

Daniel King

@danking

handling gVCF is on the medium-ish term roadmap

Where communities thrive

People

Repo info

Activity