Kipoiseq dl (#128)

* convert to kipoiseq dataloaders * remove model dataloaders * remove example files * Update model.yaml * Update model.yaml * remove model_files * use zenodo links * Update model.yaml * update pytorch models * model updates * update models * add lsgkm-SVM_redesign * use zenodo links * pwm_HOCOMOCO update * use zenodo links * update models * Delete __pycache__ * update models * remove BassetGM * ignore downloaded/ folders * update DeepBind * rename to dataloader-template * added rbp_eclip dataloaders * migrate CleTimer to zenodo * remove factornet files * fix deepTarget installation. closes kipoi/kipoi#361 * use kipoi from the master branch * fixed deepsea model loading [ci skip] * mmsplice reformat * use ignore targets [ci skip] * rename all dataloaders to kipoiseq=0.2 * fix divergent model loading * auto-download FactorNet files * add test_subset.txt * add also test_subset for mmsplice * use kipoiseq>=0.2.1 dataloader notation minor: update factornet, use * Update README.md * update uploaded DeepBind files * added test.expect to all the models * fix CleTimer output * add CpGenie/merged * CleTimer: fix model loading bug * fix the divergent unit test * fix the tensorflow version * add maxentscan fix * removed .gitatttributes * added task names to Divergent421 * remove tab * update model-template [ci-skip] * Update model.yaml * Update model.yaml * fix tests for CpGenie * test also the commmon environments * lower the prediction for labranchor and SiSp
kipoi · Nov 6, 2018 · 519a8e2 · 519a8e2
1 parent 92b23ed
commit 519a8e2
Show file tree

Hide file tree

Showing 15,411 changed files with 9,321 additions and 46,753 deletions.
diff --git a/.circleci/config.yml b/.circleci/config.yml
@@ -50,6 +50,9 @@ jobs:
       - run:
           name: run tests
           command: kipoi test-source kipoi --git-range master HEAD
+      - run:
+          name: run tests in common environmnets
+          command: kipoi test-source kipoi --git-range master HEAD --common_env
       - *save_cache
       - *store_artifacts
 
@@ -65,6 +68,9 @@ jobs:
       - run:
           name: run tests
           command: kipoi test-source kipoi --all
+      - run:
+          name: run tests in common environments
+          command: kipoi test-source kipoi --all --common_env
       - *save_cache
       - *store_artifacts
 

diff --git a/.circleci/setup.sh b/.circleci/setup.sh
@@ -44,7 +44,9 @@ model_sources:
 # ---
 
 echo "Installing Kipoi."
-pip install kipoi
+# install kipoi from the master branch for now
+pip install git+https://github.com/kipoi/kipoi
+# pip install kipoi
 
 # fi
 

diff --git a/.gitattributes b/.gitattributes
diff --git a/.gitignore b/.gitignore
@@ -105,4 +105,7 @@ ENV/
 .mypy_cache/
 *~
 
-.snakemake
+.snakemake
+
+downloaded/
+/FactorNet/FactorNet/
diff --git a/Basenji/dataloader.py b/Basenji/dataloader.py
diff --git a/Basenji/dataloader.yaml b/Basenji/dataloader.yaml
diff --git a/Basenji/example_files/hg38_chr22.fa b/Basenji/example_files/hg38_chr22.fa
diff --git a/Basenji/example_files/hg38_chr22.fa.fai b/Basenji/example_files/hg38_chr22.fa.fai
diff --git a/Basenji/example_files/intervals.bed b/Basenji/example_files/intervals.bed
diff --git a/Basenji/model.yaml b/Basenji/model.yaml
@@ -2,8 +2,19 @@ type: tensorflow
 args:
   input_nodes: "inputs"
   target_nodes: "preds"
-  checkpoint_path: "model_files/model.tf"
-  const_feed_dict_pkl: "model_files/const_feed_dict.pkl"
+  checkpoint_path:
+    meta: 
+      url: https://zenodo.org/record/1466066/files/model.tf.meta?download=1
+      md5: a679c3371ece1ac532067bdd61964025
+    index: 
+      url: https://zenodo.org/record/1466066/files/model.tf.index?download=1
+      md5: b9561a2f203a8a1589cc65183dc0898b
+    data:
+      url: https://zenodo.org/record/1466066/files/model.tf.data-00000-of-00001?download=1
+      md5: 155c2047761dbce5a18f00ba9d3fb821
+  const_feed_dict_pkl:
+    url: https://zenodo.org/record/1466066/files/const_feed_dict.pkl?download=1
+    md5: 3a76c37eb9ad255680ba774b110de1be
 # pass  
 info:
   authors:
@@ -27,17 +38,23 @@ info:
     on gene expression align well to causal variants underlying eQTLs in human populations and can
     be useful for generating mechanistic hypotheses to enable fine mapping of disease loci.
   version: 0.1
+  license: Apache License v2
   cite_as: https://doi.org/10.1101/gr.227819.117
   tags:
     - RNA expression
     - Histone modification
     - DNA accessibility
-default_dataloader: .
+default_dataloader: 
+  defined_as: kipoiseq.dataloaders.SeqIntervalDl
+  default_args:
+    auto_resize_len: 131072
+    ignore_targets: True  # don't return any output labels using the bed file
 dependencies:
   conda:
     - python=3.5
   pip:
     - tensorflow>=1.4.1
+    - kipoiseq
 schema:
   inputs:
     name: seq

diff --git a/Basenji/model_files/const_feed_dict.pkl b/Basenji/model_files/const_feed_dict.pkl
diff --git a/Basenji/model_files/model.tf-modified.meta b/Basenji/model_files/model.tf-modified.meta
diff --git a/Basenji/model_files/model.tf.data-00000-of-00001 b/Basenji/model_files/model.tf.data-00000-of-00001
diff --git a/Basenji/model_files/model.tf.index b/Basenji/model_files/model.tf.index
diff --git a/Basenji/model_files/model.tf.meta b/Basenji/model_files/model.tf.meta
diff --git a/Basenji/model_files/params.txt b/Basenji/model_files/params.txt