Merge pull request #18 from kipoi/interval_seq_rename

rename IntervalSeqDl -> SeqIntervalDl

Author: Avsecz

README.md

@@ -8,13 +8,13 @@
 
 Standard set of data-loaders for training and making predictions for DNA sequence-based models.
 
-All dataloaders in `kipoiseq.dataloaders` decorated with `@kipoi_dataloader` (IntervalSeqDl and IntervalSeqStringDl) are compatible Kipoi models and can be directly used when specifying a new model in `model.yaml`:
+All dataloaders in `kipoiseq.dataloaders` decorated with `@kipoi_dataloader` (SeqIntervalDl and StringSeqIntervalDl) are compatible Kipoi models and can be directly used when specifying a new model in `model.yaml`:
 ```yaml
 ...
 default_dataloader:
-  defined_as: kipoiseq.dataloaders.IntervalSeqDl
+  defined_as: kipoiseq.dataloaders.SeqIntervalDl
   default_args:
-    auto_resize_len: 1000 # override default args in IntervalSeqDl
+    auto_resize_len: 1000 # override default args in SeqIntervalDl
 
 dependencies:
   pip:
@@ -31,11 +31,11 @@ pip install kipoiseq
 ## Getting started
 
 ```python
-from kipoiseq.dataloaders import IntervalSeqDl
+from kipoiseq.dataloaders import SeqIntervalDl
 
-dl = IntervalSeqDl.init_example()  # use the provided example files
+dl = SeqIntervalDl.init_example()  # use the provided example files
 # your own files
-dl = IntervalSeqDl("intervals.bed", "genome.fa")
+dl = SeqIntervalDl("intervals.bed", "genome.fa")
 
 len(dl)  # length of the dataset
 
@@ -60,6 +60,6 @@ More info:
 
 ## How to write your own data-loaders
 - Read the pytorch [Data Loading and Processing Tutorial](https://pytorch.org/tutorials/beginner/data_loading_tutorial.html) to become more familiar with transforms and dataloaders
-- Read the code for `IntervalSeqDl` in [kipoiseq/dataloaders/sequence.py](https://github.com/kipoi/kipoiseq/blob/master/kipoiseq/dataloaders/sequence.py)
+- Read the code for `SeqIntervalDl` in [kipoiseq/dataloaders/sequence.py](https://github.com/kipoi/kipoiseq/blob/master/kipoiseq/dataloaders/sequence.py)
   - you can skip the `@kipoi_dataloader` and the long yaml doc-string. These are only required if you want to use dataloaders in Kipoi's model.yaml files.
 - Explore the available transforms ([functional](http://kipoi.org/kipoiseq/transforms/functional/), [class-based](http://kipoi.org/kipoiseq/transforms/transforms/)) or extractors ([kipoiseq](https://github.com/kipoi/kipoiseq/blob/master/kipoiseq/extractors.py), [genomelake](https://github.com/kundajelab/genomelake/blob/master/genomelake/extractors.py))

kipoiseq/dataloaders/sequence.py

@@ -26,7 +26,7 @@
                    Author(name='Roman Kreuzhuber', github='krrome')]
 
 # Object exported on import *
-__all__ = ['IntervalSeqDl', 'IntervalSeqStringDl', 'BedDataset']
+__all__ = ['SeqIntervalDl', 'StringSeqIntervalDl', 'BedDataset']
 
 
 class BedDataset(object):
@@ -132,7 +132,7 @@ def get_targets(self):
 
 
 @kipoi_dataloader(override={"dependencies": deps, 'info.authors': package_authors})
-class IntervalSeqStringDl(Dataset):
+class StringSeqIntervalDl(Dataset):
     """
     info:
         doc: >
@@ -258,7 +258,7 @@ def get_output_schema(cls):
 
 
 @kipoi_dataloader(override={"dependencies": deps, 'info.authors': package_authors})
-class IntervalSeqDl(Dataset):
+class SeqIntervalDl(Dataset):
     """
     info:
         doc: >
@@ -332,7 +332,7 @@ def __init__(self,
                  ignore_targets=False,
                  dtype=None):
         # core dataset, not using the one-hot encoding params
-        self.seq_dl = IntervalSeqStringDl(intervals_file, fasta_file, num_chr_fasta=num_chr_fasta,
+        self.seq_dl = StringSeqIntervalDl(intervals_file, fasta_file, num_chr_fasta=num_chr_fasta,
                                           label_dtype=label_dtype, auto_resize_len=auto_resize_len,
                                           # use_strand=use_strand,
                                           ignore_targets=ignore_targets)

notebooks/getting-started.ipynb

@@ -23,7 +23,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from kipoiseq.dataloaders import IntervalSeqDl"
+    "from kipoiseq.dataloaders import SeqIntervalDl"
    ]
   },
   {
@@ -58,7 +58,7 @@
     }
    ],
    "source": [
-    "kwargs = IntervalSeqDl.example_kwargs\n",
+    "kwargs = SeqIntervalDl.example_kwargs\n",
     "kwargs"
    ]
   },
@@ -127,7 +127,7 @@
    "outputs": [],
    "source": [
     "# setup the dataset\n",
-    "dl = IntervalSeqDl(**kwargs)"
+    "dl = SeqIntervalDl(**kwargs)"
    ]
   },
   {
@@ -252,7 +252,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "dl = IntervalSeqDl(auto_resize_len=10, **kwargs)"
+    "dl = SeqIntervalDl(auto_resize_len=10, **kwargs)"
    ]
   },
   {
@@ -708,7 +708,7 @@
    "source": [
     "### Final remarks\n",
     "\n",
-    "- See the available arguments of `IntervalSeqDl`: http://kipoi.org/kipoiseq/dataloaders/sequence/#seqdataset\n",
+    "- See the available arguments of `SeqIntervalDl`: http://kipoi.org/kipoiseq/dataloaders/sequence/#seqdataset\n",
     "- Both, the `intervals_file` and the `fasta_file` may be gzipped.\n",
     "- You may have multiple additional columns in the `intervals_file` to train a multi-task model.\n",
     "- If you are training on large datasets, find the appropriate encoding the the labels (say `bool` for binary-only labels)."

tests/dataloaders/test_sequence.py

@@ -5,7 +5,7 @@
 from pybedtools import Interval
 from kipoi.utils import override_default_kwargs
 from kipoiseq.transforms.functional import one_hot_dna
-from kipoiseq.dataloaders.sequence import IntervalSeqStringDl, IntervalSeqDl, BedDataset
+from kipoiseq.dataloaders.sequence import StringSeqIntervalDl, SeqIntervalDl, BedDataset
 
 
 @pytest.fixture
@@ -27,37 +27,37 @@ def test_min_props():
     min_set_props = ["output_schema", "type", "defined_as", "info", "args", "dependencies", "postprocessing",
                      "source", "source_dir"]
 
-    for Dl in [IntervalSeqStringDl, IntervalSeqDl]:
+    for Dl in [StringSeqIntervalDl, SeqIntervalDl]:
         props = dir(Dl)
         assert all([el in props for el in min_set_props])
 
 
 def test_fasta_based_dataset(intervals_file, fasta_file):
     # just test the functionality
-    dl = IntervalSeqStringDl(intervals_file, fasta_file)
+    dl = StringSeqIntervalDl(intervals_file, fasta_file)
     ret_val = dl[0]
     assert isinstance(ret_val["inputs"], np.ndarray)
     assert ret_val["inputs"].shape == ()
     # # test with set wrong seqlen:
-    # dl = IntervalSeqStringDl(intervals_file, fasta_file, required_seq_len=3)
+    # dl = StringSeqIntervalDl(intervals_file, fasta_file, required_seq_len=3)
     # with pytest.raises(Exception):
     #     dl[0]
 
-    dl = IntervalSeqStringDl(intervals_file, fasta_file, label_dtype="str")
+    dl = StringSeqIntervalDl(intervals_file, fasta_file, label_dtype="str")
     ret_val = dl[0]
     assert isinstance(ret_val['targets'][0], np.str_)
-    dl = IntervalSeqStringDl(intervals_file, fasta_file, label_dtype="int")
+    dl = StringSeqIntervalDl(intervals_file, fasta_file, label_dtype="int")
     ret_val = dl[0]
     assert isinstance(ret_val['targets'][0], np.int_)
-    dl = IntervalSeqStringDl(intervals_file, fasta_file, label_dtype="bool")
+    dl = StringSeqIntervalDl(intervals_file, fasta_file, label_dtype="bool")
     ret_val = dl[0]
     assert isinstance(ret_val['targets'][0], np.bool_)
     vals = dl.load_all()
     assert vals['inputs'][0] == 'GT'
 
 
 def test_seq_dataset(intervals_file, fasta_file):
-    dl = IntervalSeqDl(intervals_file, fasta_file)
+    dl = SeqIntervalDl(intervals_file, fasta_file)
     ret_val = dl[0]
 
     assert np.all(ret_val['inputs'] == one_hot_dna("GT"))
@@ -67,7 +67,7 @@ def test_seq_dataset(intervals_file, fasta_file):
 
 @pytest.fixture
 def example_kwargs():
-    return IntervalSeqDl.example_kwargs
+    return SeqIntervalDl.example_kwargs
 
 
 @pytest.mark.parametrize("alphabet_axis", list(range(0, 4)))
@@ -87,10 +87,10 @@ def test_seq_dataset_reshape(alphabet_axis, dummy_axis, example_kwargs):
     if (alphabet_axis == dummy_axis_int) or (alphabet_axis == -1) or (dummy_axis_int == -1) or \
             (alphabet_axis >= 3) or (dummy_axis_int >= 3) or ((alphabet_axis >= 2) and (dummy_axis is None)):
         with pytest.raises(Exception):
-            seq_dataset = IntervalSeqDl(**kwargs)
+            seq_dataset = SeqIntervalDl(**kwargs)
         return None
 
-    seq_dataset = IntervalSeqDl(**kwargs)
+    seq_dataset = SeqIntervalDl(**kwargs)
 
     # test the single sample works
     reshaped = seq_dataset[0]['inputs']
@@ -104,7 +104,7 @@ def test_seq_dataset_reshape(alphabet_axis, dummy_axis, example_kwargs):
 
 
 # download example files
-@pytest.mark.parametrize("cls", [IntervalSeqStringDl, IntervalSeqDl])
+@pytest.mark.parametrize("cls", [StringSeqIntervalDl, SeqIntervalDl])
 def test_examples_exist(cls):
     ex = cls.init_example()
     example_kwargs = cls.example_kwargs
@@ -120,7 +120,7 @@ def test_examples_exist(cls):
 
 
 def test_output_schape():
-    Dl = deepcopy(IntervalSeqDl)
+    Dl = deepcopy(SeqIntervalDl)
     assert Dl.get_output_schema().inputs.shape == (None, 4)
     override_default_kwargs(Dl, {"auto_resize_len": 100})
     assert Dl.get_output_schema().inputs.shape == (100, 4)

tests/dont_test_4_integration.py

@@ -3,7 +3,7 @@
 import kipoi_veff.snv_predict as sp
 import pytest
 from kipoi_veff.utils.generic import ModelInfoExtractor
-from kipoiseq.dataloaders.sequence import IntervalSeqDl
+from kipoiseq.dataloaders.sequence import SeqIntervalDl
 import os
 from kipoi.pipeline import install_model_requirements
 
@@ -12,7 +12,7 @@
 
 def test_deepsea():
     model = kipoi.get_model("DeepSEA/variantEffects")
-    mie = ModelInfoExtractor(model, IntervalSeqDl)
+    mie = ModelInfoExtractor(model, SeqIntervalDl)
 
 
 def test_var_eff_pred_varseq(tmpdir):
@@ -22,7 +22,7 @@ def test_var_eff_pred_varseq(tmpdir):
     #
     model = kipoi.get_model(model_name, source="kipoi")
     # The preprocessor
-    Dataloader = IntervalSeqDl
+    Dataloader = SeqIntervalDl
     #
     dataloader_arguments = {"intervals_file": "example_files/intervals.bed",
                             "fasta_file": "example_files/hg38_chr22.fa",