Merge pull request #15 from 4dn-dcic/dev

Dev
4dn-dcic · Dec 6, 2017 · b6838d8 · b6838d8
2 parents f144ca1 + ba9952a
commit b6838d8
Show file tree

Hide file tree

Showing 16 changed files with 282 additions and 169 deletions.
diff --git a/Benchmark/Benchmark.py b/Benchmark/Benchmark.py
@@ -54,6 +54,8 @@ def benchmark(app_name, input_json, raise_error=False):
         return(hi_c_processing_partc(input_json))
     elif app_name == 'pairs-patch':
         return(pairs_patch(input_json))
+    elif app_name == 'repliseq-parta':
+        return(repliseq_parta(input_json))
     else:
         if raise_error:
             raise AppNameUnavailableException
@@ -327,6 +329,43 @@ def hi_c_processing_partc(input_json):
     return(r.as_dict())
 
 
+def repliseq_parta(input_json):
+    assert 'input_size_in_bytes' in input_json
+    assert 'fastq' in input_json.get('input_size_in_bytes')
+    assert 'bwaIndex' in input_json.get('input_size_in_bytes')
+
+    # cpu
+    nthreads = 4  # default from cwl
+    if 'parameters' in input_json:
+        if 'nthreads' in input_json.get('parameters'):
+            nthreads = input_json.get('parameters').get('nthreads')
+
+    # space
+    input_sizes = input_json.get('input_size_in_bytes')
+    data_input_size = input_sizes.get('fastq')
+    total_input_size = data_input_size + input_sizes.get('bwaIndex')
+    output_bam_size = data_input_size * 2
+    output_clipped_fq_size = data_input_size
+    output_size = output_bam_size * 3 + output_clipped_fq_size
+    intermediate_index_size = input_sizes.get('bwaIndex') * 2
+    copied_input_size = data_input_size * 7  # copied and unzipped
+    total_intermediate_size \
+        = intermediate_index_size + output_size + copied_input_size
+    total_output_size = output_size
+    additional_size_in_gb = 10
+
+    total_file_size_in_bp \
+        = total_input_size + total_intermediate_size + total_output_size
+    total_size = total_file_size_in_bp / GB_IN_BYTES + additional_size_in_gb
+
+    # mem
+    mem = input_sizes.get('bwaIndex') * 4 / MB_IN_BYTES + (nthreads * 500)
+
+    r = BenchmarkResult(size=total_size, mem=mem, cpu=nthreads)
+
+    return(r.as_dict())
+
+
 def get_aws_ec2_info_file():
     this_dir, _ = os.path.split(__file__)
     return(os.path.join(this_dir, "aws", "Amazon EC2 Instance Comparison.csv"))

diff --git a/Benchmark/__pycache__/Benchmark.cpython-35.pyc b/Benchmark/__pycache__/Benchmark.cpython-35.pyc
diff --git a/Benchmark/_version.py b/Benchmark/_version.py
@@ -1,4 +1,4 @@
 """Version information."""
 
 # The following line *must* be the last in the module, exactly as formatted:
-__version__ = "0.2.0"
+__version__ = "0.2.3"
diff --git a/cwl_awsem/repliseq/align_se.cwl → cwl_awsem/repliseq/align.cwl b/cwl_awsem/repliseq/align_se.cwl → cwl_awsem/repliseq/align.cwl
@@ -1,7 +1,7 @@
 {
     "hints": [
         {
-            "dockerPull": "duplexa/4dn-repliseq:v9",
+            "dockerPull": "duplexa/4dn-repliseq:v10",
             "class": "DockerRequirement"
         }
     ],
@@ -41,7 +41,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#prefix",
@@ -53,7 +53,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outdir",
@@ -67,7 +67,7 @@
     "outputs": [
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_bam",
@@ -77,7 +77,7 @@
         }
     ],
     "baseCommand": [
-        "run-align-se.sh"
+        "run-align.sh"
     ],
     "requirements": [
         {

diff --git a/cwl_awsem/repliseq/align_pe.cwl b/cwl_awsem/repliseq/align_pe.cwl
diff --git a/cwl_awsem/repliseq/clip.cwl b/cwl_awsem/repliseq/clip.cwl
@@ -1,7 +1,7 @@
 {
     "hints": [
         {
-            "dockerPull": "duplexa/4dn-repliseq:v9",
+            "dockerPull": "duplexa/4dn-repliseq:v10",
             "class": "DockerRequirement"
         }
     ],
@@ -20,7 +20,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outprefix",
@@ -32,7 +32,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outdir",
@@ -46,7 +46,7 @@
     "outputs": [
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_clipped_fastq",
@@ -56,7 +56,7 @@
         },
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_log",

diff --git a/cwl_awsem/repliseq/count.cwl b/cwl_awsem/repliseq/count.cwl
@@ -1,7 +1,7 @@
 {
     "hints": [
         {
-            "dockerPull": "duplexa/4dn-repliseq:v9",
+            "dockerPull": "duplexa/4dn-repliseq:v10",
             "class": "DockerRequirement"
         }
     ],
@@ -20,7 +20,7 @@
         },
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "inputBinding": {
@@ -42,18 +42,7 @@
         },
         {
             "type": [
-                "int"
-            ],
-            "id": "#minsize",
-            "inputBinding": {
-                "position": 6,
-                "separate": true
-            },
-            "default": 10000000
-        },
-        {
-            "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outprefix",
@@ -65,7 +54,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outdir",
@@ -79,7 +68,7 @@
     "outputs": [
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_count_bg",

diff --git a/cwl_awsem/repliseq/dedup.cwl b/cwl_awsem/repliseq/dedup.cwl
@@ -1,7 +1,7 @@
 {
     "hints": [
         {
-            "dockerPull": "duplexa/4dn-repliseq:v9",
+            "dockerPull": "duplexa/4dn-repliseq:v10",
             "class": "DockerRequirement"
         }
     ],
@@ -20,7 +20,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outprefix",
@@ -32,7 +32,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outdir",
@@ -46,7 +46,7 @@
     "outputs": [
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_deduped_bam",
@@ -56,7 +56,7 @@
         },
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_log",

diff --git a/cwl_awsem/repliseq/filter.cwl b/cwl_awsem/repliseq/filter.cwl
@@ -1,7 +1,7 @@
 {
     "hints": [
         {
-            "dockerPull": "duplexa/4dn-repliseq:v9",
+            "dockerPull": "duplexa/4dn-repliseq:v10",
             "class": "DockerRequirement"
         }
     ],
@@ -30,7 +30,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outprefix",
@@ -42,7 +42,7 @@
         },
         {
             "type": [
-                "null",
+
                 "string"
             ],
             "id": "#outdir",
@@ -56,7 +56,7 @@
     "outputs": [
         {
             "type": [
-                "null",
+
                 "File"
             ],
             "id": "#out_filtered_bg",