def test_and_or():
df = tibble(x=1, y=2, z=3, w=4)
out = df >> select(c(f.x, f.y) & c(f.y, f.z))
assert out.columns.tolist() == ["y"]
out = df >> mutate(a=f.x & f.y)
assert out.a.tolist() == [True]
out = df >> mutate(a=True & f.y)
assert out.a.tolist() == [True]
out = df >> select(c(f.x, f.y) | c(f.y, f.z))
assert out.columns.tolist() == ["x", "y", "z"]
def test_inv():
df = tibble(x=1, y=2)
out = df >> select(~f.x)
assert out.columns.tolist() == ["y"]
df = tibble(x=True)
out = df >> mutate(y=~f.x)
assert out.y.tolist() == [False]
def gene_name_conversion(
genes,
species,
infmt,
outfmt,
notfound,
):
"""Convert gene names using MyGeneInfo
Args:
genes: A sequence of genes
species: The species to limit the query
Supported: human, mouse, rat, fruitfly, nematode, zebrafish,
thale-cress, frog and pig
infmt: What's the original gene name format
Available fields
https://docs.mygene.info/en/latest/doc/query_service.html#available-fields
outfmt: What's the target gene name format
notfound: What to do if a conversion cannot be done.
use-query: Ignore the conversion and use the original name
skip: Ignore the conversion and skip the entire row in input file
error: Report error
Returns:
A dataframe with two columns, query and `outfmt`.
"""
out = (mygene.querymany(
genes,
scopes=infmt,
fields=outfmt,
as_dataframe=True,
df_index=False,
species=species,
) >> group_by(f.query) >> arrange(desc(f._score)) >> slice_head(1) >>
select(~c(f._id, f._score, f.notfound)))
if isinstance(outfmt, str):
outfmt = [of.strip() for of in outfmt.split(",")]
out = tibble(query=genes) >> left_join(out, by=f.query)
if notfound == "use-query":
out = out >> mutate(
across(
outfmt,
lambda col, query: if_else(is_na(col), query, col),
query=f.query,
))
elif notfound == "error" and any(is_na(out[outfmt[0]])):
nagenes = out >> filter(is_na(f[outfmt[0]])) >> pull(f.query)
raise QueryGenesNotFound(nagenes)
elif notfound == "skip":
out = out >> filter(~is_na(f[outfmt[0]]))
return out
from pipen import Proc
from biopipen.ns.web import Download
from biopipen.ns.misc import Str2File
from biopipen.ns.gsea import FGSEA
from datar.all import flatten, tibble, select
from biopipen.core.testing import get_pipeline
FGSEA = Proc.from_proc(
FGSEA,
requires=[Download, Str2File],
input_data=lambda ch1, ch2: tibble(
*flatten(ch1),
ch2,
_name_repair="minimal",
) >> select(0, 2, 1),
envs={"clscol": "Group", "classes": ["MMP9", "CTRL"]}
)
def pipeline():
return get_pipeline(__file__).set_starts(Download, Str2File).set_data(
[
"https://www.ncbi.nlm.nih.gov/geo/download/"
"?acc=GSE179367"
"&format=file"
"&file=GSE179367%5Fgene%5Fcount%2Ereal%2Etxt%2Egz",
"https://www.genepattern.org/tutorial/linkedFiles/"
"export_gnf.GENE_SYMBOL.gmt",
],
[(
def test_neg():
df = tibble(x=1, y=2)
out = df >> select(-f[:1])
assert out.columns.tolist() == ["y"]
def test_or_():
df = tibble(x=1, y=2, z=3)
out = df >> select(c(f.x, f.y) | [f.y, f.z])
assert_frame_equal(out, tibble(x=1, y=2, z=3))
outfmt=outfmt,
notfound=notfound,
)
converted.columns = [colname] + converted.columns[1:].tolist()
if output == "only":
out = converted
elif output == "keep":
out = df >> right_join(converted, by=colname, suffix=["", "_converted"])
elif output == "drop":
out = df >> right_join(
converted,
by=colname, suffix=["", "_converted"]
) >> select(~c(colname))
elif output == "replace":
out = df >> right_join(
converted, by=colname, suffix=["", "_converted"]
)
converted_cols = out.columns[-len(converted.columns)+1:].tolist()
pos = df.columns.get_indexer([colname])[0]
out = out >> relocate(
converted_cols, _after=pos+1
) >> select(~c(colname))
else:
raise ValueError(f"Unknown output mode: {output}.")
out.to_csv(outfile, **outopts)