def test_Zipdist2():
class X():
def __init__(self, name):
self.name = name
x = X(name='examplezip2')
x.year = 2020
x.exnp = np.zeros(3)
x.expd = pd.DataFrame({"A": [1, 2, 3], "B": [2, 4, 6]})
z = Zipdist2(name="examplezip2", target=x)
assert z.target is x
z._save(dest="xxx", dest_tar="xxx.tar.gz")
assert os.path.isfile("xxx.tar.gz")
x2 = X(name='examplezip2')
assert 'year' not in x2.__dict__.keys()
z._build(target=x2, dest="xxx", dest_tar="xxx.tar.gz")
print(x.__dict__.keys())
assert isinstance(x2.exnp, np.ndarray)
#assert np.all(x2.exnp == x.exnp)
print(x2.exnp)
print(x.exnp)
assert np.array_equal(x2.exnp, x.exnp)
assert np.all(x2.expd.equals(x.expd))
assert np.all(x2.name == x.name)
assert np.all(x2.year == x.year)
def test_Zipdist2_example():
from zipdist.zip2 import Zipdist2
import pandas as pd
import numpy as np
import os
class X():
def __init__(self, name):
self.name = name
x = X(name='example_target_object')
x.example_simple_attr = [1989, 2020]
x.example_np_attr = np.array([[0, 1, 2, 3], [4, 5, 6, 7]])
x.example_pd_attr = pd.DataFrame({"A": [1, 2, 3], "B": [2, 4, 6]})
z = Zipdist2(name="zipper", target=x)
z._save(dest="archive", dest_tar="archive.tar.gz")
assert os.path.isfile("archive.tar.gz")
x_new = X(name='example_target_object')
z = Zipdist2(name="zipper", target=x_new)
# You can use the z._build() and reload all object attributes
z._build(target=x_new, dest="archive", dest_tar="archive.tar.gz")
print(f"For instance:")
print(f"\tx_new.example_simple_attr:\n{x_new.example_simple_attr}")
print(f"\tx_new.example_np_attr:\n{x_new.example_np_attr}")
print(f"\tx_new.example_pd_attr:\n{x_new.example_pd_attr}")
# You can alternatively use the z._ready() and reload object attributes one by one
x_new = X(name='example_target_object')
z = Zipdist2(name="zipper", target=x_new)
z._ready(target=x_new, dest="archive", dest_tar="archive.tar.gz")
z._reload_complex(k='example_np_attr')
z._reload_simple(k='example_simple_attr')
def test_TCRrep_func_tcrdist2_save_manual_rebuild(chains = ['beta','alpha'], metric = "nw"):
cpu = multiprocessing.cpu_count()
# really basic example
df = pd.read_csv(opj('tcrdist', 'datasets', 'dash.csv'))
df = df[df.epitope.isin(['NP'])]
tr = TCRrep(cell_df=df, chains=chains, organism='mouse')
tr.tcrdist2(processes = cpu, metric = metric, dump = True, reduce = True, save=True, dest = "default_archive", dest_tar_name = "default_archive.tar.gz" )
# Cleanup folder that you just made
os.system("rm -rf myTCRrep_archive")
# Rebuild
tr = TCRrep(cell_df=df.iloc[0:0,:], chains=chains, organism='mouse')
z = Zipdist2(name = "test_only", target = tr)
z._build(dest_tar = "default_archive.tar.gz", target = tr)
assert isinstance(tr.paired_tcrdist, np.ndarray )
assert isinstance(tr.pw_tcrdist, np.ndarray )
assert np.array_equal(tr.pw_tcrdist, tr.paired_tcrdist)
def archive(self,
dest=None,
dest_tar_name=None,
verbose=True,
use_csv=False):
"""
Use Zipdist2 to Make an Archive.tar.gz
Parameters
----------
dest : str
e.g., 'default_archive'
dest_tar_name : str
e.g., 'default_archive.tar.gz'
verbose : bool
if True, report steps in archive process
use_csv : bool
if True, archive will include .csv file. Useful for porting files to other applications, but creates large files.
Example
-------
.. code-block:: python
tr = TCRrep(cell_df = pd.DataFrame(), organism = "mouse")
tr.archive(dest = "default_archive", dest_tar_name = "default_archive.tar.gz")
Notes
-----
See :py:meth:`tcrdist.repertoire.rebuild`: for reubilding a TCRrep instance from
an TCRrep archive .tar.gz file.
"""
if dest is None:
dest = self.archive_name
if dest_tar_name is None:
dest_tar_name = f"{dest}.tar.gz"
self.cell_df_index = self.cell_df.index.copy()
self.cell_df = self.cell_df.reset_index()
z = Zipdist2(name=dest_tar_name, target=self)
z._save(dest=dest,
dest_tar=dest_tar_name,
verbose=verbose,
use_csv=use_csv)
sys.stdout.write(
f"\tArchiving your TCRrep using Zipdist2 in [{dest_tar_name}]\n")
def make_tar_gz():
class X():
def __init__(self, name):
self.name = name
x = X(name='xxx')
x.year = 2020
x.exnp = np.zeros(3)
x.exnp1 = np.zeros(10) #, "np.zeros(10)"
x.exnp2 = np.ones(10) #, "np.ones(10)"),
x.exnp3 = np.array([[0, 1, 2, 3], [4, 5, 6, 7]]) #, "2D np.array"),
x.exnp4 = np.random.randint(0, 100, size=(30, 10, 2)) #, "3D np.array") ]
x.expd = pd.DataFrame({"A": [1, 2, 3], "B": [2, 4, 6]})
z = Zipdist2(name="xxx", target=x)
z._save(dest="xxx", dest_tar="xxx.tar.gz")
os.system('rm -rf xxx')
return x
def rebuild(self, dest=None, dest_tar_name=None, verbose=True):
"""
Use Zipdist2 to reubild a TCRrep instance from an Archive.tar.gz
Parameters
----------
dest_tar_name : str
e.g., 'default_archive.tar.gz'
verbose : bool
If True, report rebuilding process steps.
Example
-------
Shows :py:meth:`tcrdist.repertoire.archive` and :py:meth:`tcrdist.repertoire.rebuild`
used together.
.. code-block:: python
tr = TCRrep(cell_df = pd.DataFrame(), organism = "mouse")
tr.archive(dest = "default_archive", dest_tar_name = "default_archive.tar.gz")
tr_new = TCRrep(cell_df = pd.DataFrame(), organism = "mouse")
tr_new.rebuild(dest_tar_name = "default_archive.tar.gz")
Notes
-----
See :py:meth:`tcrdist.repertoire.archive` for creating TCRrep archive file.
"""
#tr = TCRrep(cell_df=df.iloc[0:0,:], chains=chains, organism='mouse')
if dest is None:
dest = self.archive_name
if dest_tar_name is None:
dest_tar_name = f"{dest}.tar.gz"
z = Zipdist2(name=dest, target=self)
z._build(dest_tar=dest_tar_name, target=self, verbose=verbose)
# VALIDATION OF INPUTS
# check that chains are valid.
self._validate_organism()
self._validate_chains()
# check that is a pd.DataFrame
self._validate_cell_df()
# RE INIT the REFERENCE DB see repertoire_db.py
self.all_genes = repertoire_db.RefGeneSet(self.db_file).all_genes