Mercurial > repos > saket-choudhary > sift_web

changeset 5:697e66764209 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author saket-choudhary
date Tue, 07 Oct 2014 19:19:11 -0400
parents 5cc3210369ac
children
files sift_web/sift_web.py sift_web/sift_web.xml sift_web/test-data/sift_input.txt sift_web/test-data/sift_output_condensed.tsv sift_web/test-data/sift_output_full.tsv sift_web/test-data/sift_output_summary.tsv
diffstat 6 files changed, 128 insertions(+), 34 deletions(-) [+]
line wrap: on
line diff
--- a/sift_web/sift_web.py	Thu Sep 04 17:46:51 2014 -0400
+++ b/sift_web/sift_web.py	Tue Oct 07 19:19:11 2014 -0400
@@ -15,26 +15,7 @@
 
 
 def retry(ExceptionToCheck, tries=10, delay=3, backoff=2, logger=None):
-    """Retry calling the decorated function using an exponential backoff.
-
-    http://www.saltycrane.com/blog/2009/11/trying-out-retry-decorator-python/
-    original from: http://wiki.python.org/moin/PythonDecoratorLibrary#Retry
-
-    :param ExceptionToCheck: the exception to check. may be a tuple of
-        exceptions to check
-    :type ExceptionToCheck: Exception or tuple
-    :param tries: number of times to try (not retry) before giving up
-    :type tries: int
-    :param delay: initial delay between retries in seconds
-    :type delay: int
-    :param backoff: backoff multiplier e.g. value of 2 will double the delay
-        each retry
-    :type backoff: int
-    :param logger: logger to use. If None, print
-    :type logger: logging.Logger instance
-    """
     def deco_retry(f):
-
         @wraps(f)
         def f_retry(*args, **kwargs):
             mtries, mdelay = tries, delay
@@ -53,9 +34,7 @@
                     mtries -= 1
                     mdelay *= backoff
             return f(*args, **kwargs)
-
         return f_retry  # true decorator
-
     return deco_retry
 
 
@@ -68,6 +47,7 @@
         self.url_dict = {'full': self.full_download_url,
                          'condensed': self.condensed_download_url,
                          'summary': self.summary_download_url}
+        self.job_id = None
 
     def upload(self, inputpath):
         payload = {'table': 'human37_66'}
@@ -77,19 +57,19 @@
         return request.text
 
     @retry(requests.exceptions.HTTPError)
-    def get_full_data(self, job_id, full_output):
+    def get_full_data(self, full_output):
         r = requests.request(
-            'GET', (self.full_download_url) % (job_id, job_id))
+            'GET', (self.full_download_url) % (self.job_id, self.job_id))
         if r.text != 'No file exists':
             with open(full_output, 'wb') as f:
                 f.write(r.text)
         else:
-            return requests.HTTPError()
+            raise(requests.HTTPError())
 
     @retry(requests.exceptions.HTTPError)
-    def get_condensed_data(self, job_id, condensed_output):
+    def get_condensed_data(self, condensed_output):
         r = requests.request(
-            'GET', (self.condensed_download_url) % (job_id, job_id))
+            'GET', (self.condensed_download_url) % (self.job_id, self.job_id))
         if r.text != 'No file exists':
             with open(condensed_output, 'wb') as f:
                 f.write(r.text)
@@ -97,9 +77,9 @@
             raise(requests.HTTPError())
 
     @retry(requests.exceptions.HTTPError)
-    def get_summary_data(self, job_id, summary_output):
+    def get_summary_data(self, summary_output):
         r = requests.request(
-            'GET', (self.summary_download_url) % (job_id, job_id))
+            'GET', (self.summary_download_url) % (self.job_id, self.job_id))
         if r.text != 'No file exists':
             with open(summary_output, 'wb') as f:
                 f.write(r.text)
@@ -119,9 +99,10 @@
     soup = BeautifulSoup(content)
     p = soup.findAll('p')
     job_id = p[1].string.split(':')[-1].replace(' ', '').replace(').', '')
-    sift_web.get_full_data(job_id, args.output1)
-    sift_web.get_condensed_data(job_id, args.output2)
-    sift_web.get_summary_data(job_id, args.output3)
+    sift_web.job_id = job_id
+    sift_web.get_full_data(args.output1)
+    sift_web.get_condensed_data(args.output2)
+    sift_web.get_summary_data(args.output3)
 
 
 if __name__ == '__main__':
--- a/sift_web/sift_web.xml	Thu Sep 04 17:46:51 2014 -0400
+++ b/sift_web/sift_web.xml	Tue Oct 07 19:19:11 2014 -0400
@@ -13,9 +13,17 @@
         <param name="input" format="txt" type="data" label="Input Variants"/>
     </inputs>
     <outputs>
-        <data name="output1" format="tabular"/>
-        <data name="output2" format="tabular"/>
-        <data name="output3" format="tabular"/>
+        <data name="output1" format="tabular" label="${tool.name} on ${on_string}: Full Output"/>
+        <data name="output2" format="tabular" label="${tool.name} on ${on_string}: Condensed Output"/>
+        <data name="output3" format="tabular" label="${tool.name} on ${on_string}: Summary table"/>
     </outputs>
+    <tests>
+        <test>
+            <param name="input" value="sift_input.txt"/>
+            <output name="output1" file="sift_output_full.tsv"/>
+            <output name="output2" file="sift_output_condensed.tsv"/>
+            <output name="output3" file="sift_output_summary.tsv"/>
+        </test>
+    </tests>
 </tool>
 
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/sift_web/test-data/sift_input.txt	Tue Oct 07 19:19:11 2014 -0400
@@ -0,0 +1,14 @@
+1,100382265,C,G,user comment 1
+1,100380997,A,G,user comment 2
+22,30163533,A,C
+X,12905093,A,T
+2,230633386,G,C
+1,100382265,C,A
+7,117199641,ATCA,.
+7,117199647,TTT,.
+10,50184923,TGG,.
+12,121438957,ACC,.
+1,43217995,G,GCCA
+10,102762472,G,GGCG
+9,117856130,T,G
+9,117856135,C,G
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/sift_web/test-data/sift_output_condensed.tsv	Tue Oct 07 19:19:11 2014 -0400
@@ -0,0 +1,15 @@
+#ROW_NO.	INPUT	PROTEIN_ID	LENGTH	STRAND	CODON_CHANGE	POS	RESIDUE_REF	RESIDUE_ALT	TYPE	SCORE	PREDICTION (cutoff=-2.5)	#SEQ	#CLUSTER	SCORE	PREDICTION (cutoff=0.05)	MEDIAN_INFO	#SEQ	dbSNP_ID
+1	1,100382265,C,G,user comment 1	ENSP00000294724	1532	1	TCC [C/G]GA CAT	1487	R	G	Single AA Change	-1.75	Neutral	107	30	0.118	Tolerated	2.82	71	rs12118058
+2	1,100380997,A,G,user comment 2	ENSP00000294724	1532	1	GCA G[A/G]A AAA	1405	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.80	72	rs28730708
+3	22,30163533,A,C	ENSP00000333278	132	1	ACG [A/C]GG GGA	49	R	R	Synonymous	0.00	Neutral	3	1	NA	NA	NA	NA	
+4	X,12905093,A,T	ENSP00000370034	1049	1	AAT G[A/T]A AGC	489	E	V	Single AA Change	-1.99	Neutral	265	30	0.044	Damaging	2.87	154	
+5	2,230633386,G,C	ENSP00000373696	2040	-1	GAG [C/G]AG CAG	1958	Q	E	Single AA Change	0.46	Neutral	127	30	0.602	Tolerated	2.89	136	
+6	1,100382265,C,A	ENSP00000294724	1532	1	TCC [C/A]GA CAT	1487	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.82	71	
+7	7,117199641,ATCA,.	ENSP00000003084	1480	1					Frameshift	NA	NA	273	30	NA	NA	NA	NA	
+8	7,117199647,TTT,.	ENSP00000003084	1480	1	ATC [TTT/-] GGT	508	F	.	Deletion	-13.79	Deleterious	273	30	NA	NA	NA	NA	rs121909001
+9	10,50184923,TGG,.	ENSP00000320563	3184	1	GCC [TGG/-] GGC	3064	W	.	Deletion	-6.39	Deleterious	101	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000443112	662	1	GAG [ACC/-] TTC	651	T	.	Deletion	-3.66	Deleterious	158	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000236040	804	-1	CCA [-/TGG] CAT	506	P	PW	Insertion	-9.11	Deleterious	157	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000359240	669	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.26	Deleterious	136	30	NA	NA	NA	NA	
+13	9,117856130,T,G	record not found																rs1513
+14	9,117856135,C,G	record not found																
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/sift_web/test-data/sift_output_full.tsv	Tue Oct 07 19:19:11 2014 -0400
@@ -0,0 +1,63 @@
+#ROW_NO.	INPUT	PROTEIN_ID	LENGTH	STRAND	CODON_CHANGE	POS	RESIDUE_REF	RESIDUE_ALT	TYPE	SCORE	PREDICTION (cutoff=-2.5)	#SEQ	#CLUSTER	SCORE	PREDICTION (cutoff=0.05)	MEDIAN_INFO	#SEQ	dbSNP_ID
+1	1,100382265,C,G,user comment 1	ENSP00000294724	1532	1	TCC [C/G]GA CAT	1487	R	G	Single AA Change	-1.75	Neutral	107	30	0.118	Tolerated	2.82	71	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000354635	1515	1	TCC [C/G]GA CAT	1470	R	G	Single AA Change	-1.75	Neutral	107	30	0.114	Tolerated	2.84	71	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000354971	1516	1	TCC [C/G]GA CAT	1471	R	G	Single AA Change	-1.75	Neutral	107	30	0.108	Tolerated	2.83	73	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000355106	1532	1	TCC [C/G]GA CAT	1487	R	G	Single AA Change	-1.75	Neutral	107	30	0.118	Tolerated	2.82	71	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000359180	1516	1	TCC [C/G]GA CAT	1471	R	G	Single AA Change	-1.75	Neutral	107	30	0.108	Tolerated	2.83	73	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000359182	1532	1	TCC [C/G]GA CAT	1487	R	G	Single AA Change	-1.75	Neutral	107	30	0.118	Tolerated	2.82	71	rs12118058
+1	1,100382265,C,G,user comment 1	ENSP00000359184	1532	1	TCC [C/G]GA CAT	1487	R	G	Single AA Change	-1.75	Neutral	107	30	0.118	Tolerated	2.82	71	rs12118058
+2	1,100380997,A,G,user comment 2	ENSP00000294724	1532	1	GCA G[A/G]A AAA	1405	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.80	72	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000354635	1515	1	GCA G[A/G]A AAA	1388	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.82	72	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000354971	1516	1	GCA G[A/G]A AAA	1389	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.81	74	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000355106	1532	1	GCA G[A/G]A AAA	1405	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.80	72	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000359180	1516	1	GCA G[A/G]A AAA	1389	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.81	74	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000359182	1532	1	GCA G[A/G]A AAA	1405	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.80	72	rs28730708
+2	1,100380997,A,G,user comment 2	ENSP00000359184	1532	1	GCA G[A/G]A AAA	1405	E	G	Single AA Change	-4.58	Deleterious	107	30	0.004	Damaging	2.80	72	rs28730708
+3	22,30163533,A,C	ENSP00000332887	63	1	AAC G[A/C]G GGG	49	E	A	Single AA Change	-3.11	Deleterious	87	30	0.020	Damaging	2.79	85	
+3	22,30163533,A,C	ENSP00000333278	132	1	ACG [A/C]GG GGA	49	R	R	Synonymous	0.00	Neutral	3	1	NA	NA	NA	NA	
+3	22,30163533,A,C	ENSP00000384962	62	1	AAC G[A/C]G GGG	49	E	A	Single AA Change	-3.05	Deleterious	85	30	0.057	Tolerated	3.02	57	
+3	22,30163533,A,C	ENSP00000385696	82	1	AAC G[A/C]G GGG	49	E	A	Single AA Change	-3.00	Deleterious	88	30	0.036	Damaging	3.07	83	
+4	X,12905093,A,T	ENSP00000370034	1049	1	AAT G[A/T]A AGC	489	E	V	Single AA Change	-1.99	Neutral	265	30	0.044	Damaging	2.87	154	
+5	2,230633386,G,C	ENSP00000283943	1992	-1	GAG [C/G]AG CAG	1910	Q	E	Single AA Change	0.46	Neutral	127	30	0.811	Tolerated	2.90	99	
+5	2,230633386,G,C	ENSP00000373696	2040	-1	GAG [C/G]AG CAG	1958	Q	E	Single AA Change	0.46	Neutral	127	30	0.602	Tolerated	2.89	136	
+5	2,230633386,G,C	ENSP00000373697	1722	-1	GAG [C/G]AG CAG	1640	Q	E	Single AA Change	0.33	Neutral	129	30	1.000	Tolerated	2.96	95	
+5	2,230633386,G,C	ENSP00000408330	187	-1	GAG [C/G]AG CAG	170	Q	E	Single AA Change	0.86	Neutral	120	30	1.000	Tolerated	2.82	62	
+6	1,100382265,C,A	ENSP00000294724	1532	1	TCC [C/A]GA CAT	1487	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.82	71	
+6	1,100382265,C,A	ENSP00000354635	1515	1	TCC [C/A]GA CAT	1470	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.84	71	
+6	1,100382265,C,A	ENSP00000354971	1516	1	TCC [C/A]GA CAT	1471	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.83	73	
+6	1,100382265,C,A	ENSP00000355106	1532	1	TCC [C/A]GA CAT	1487	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.82	71	
+6	1,100382265,C,A	ENSP00000359180	1516	1	TCC [C/A]GA CAT	1471	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.83	73	
+6	1,100382265,C,A	ENSP00000359182	1532	1	TCC [C/A]GA CAT	1487	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.82	71	
+6	1,100382265,C,A	ENSP00000359184	1532	1	TCC [C/A]GA CAT	1487	R	R	Synonymous	0.00	Neutral	107	30	1.000	Tolerated	2.82	71	
+7	7,117199641,ATCA,.	ENSP00000003084	1480	1					Frameshift	NA	NA	273	30	NA	NA	NA	NA	
+7	7,117199641,ATCA,.	ENSP00000389119	1438	1					Frameshift	NA	NA	273	30	NA	NA	NA	NA	
+7	7,117199641,ATCA,.	ENSP00000403677	1419	1					Frameshift	NA	NA	275	30	NA	NA	NA	NA	
+8	7,117199647,TTT,.	ENSP00000003084	1480	1	ATC [TTT/-] GGT	508	F	.	Deletion	-13.79	Deleterious	273	30	NA	NA	NA	NA	rs121909001
+8	7,117199647,TTT,.	ENSP00000389119	1438	1	ATC [TTT/-] GGT	478	F	.	Deletion	-13.79	Deleterious	273	30	NA	NA	NA	NA	rs121909001
+8	7,117199647,TTT,.	ENSP00000403677	1419	1	ATC [TTT/-] GGT	447	F	.	Deletion	-14.12	Deleterious	275	30	NA	NA	NA	NA	rs121909001
+9	10,50184923,TGG,.	ENSP00000265453	1271	1	GCC [TGG/-] GGC	1151	W	.	Deletion	-8.05	Deleterious	103	30	NA	NA	NA	NA	
+9	10,50184923,TGG,.	ENSP00000309314	2267	1	GCC [TGG/-] GGC	2155	W	.	Deletion	-7.09	Deleterious	103	30	NA	NA	NA	NA	
+9	10,50184923,TGG,.	ENSP00000320563	3184	1	GCC [TGG/-] GGC	3064	W	.	Deletion	-6.39	Deleterious	101	30	NA	NA	NA	NA	
+9	10,50184923,TGG,.	ENSP00000409696	3176	1	GCC [TGG/-] GGC	3064	W	.	Deletion	-6.39	Deleterious	101	30	NA	NA	NA	NA	
+9	10,50184923,TGG,.	ENSP00000443499	624	1	GCC [TGG/-] GGC	527	W	.	Deletion	-8.09	Deleterious	110	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000257555	631	1	GAG [ACC/-] TTC	620	T	.	Deletion	-3.40	Deleterious	159	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000339938	572	1	GAG [ACC/-] TTC	561	T	.	Deletion	-3.47	Deleterious	163	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000342683	521	1	GAG [ACC/-] TTC	510	T	.	Deletion	-2.88	Deleterious	164	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000438300	523	1	GAG [ACC/-] TTC	512	T	.	Deletion	-3.05	Deleterious	164	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000438531	452	1	GAG [ACC/-] TTC	441	T	.	Deletion	-3.49	Deleterious	165	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000438804	638	1	GAG [ACC/-] TTC	627	T	.	Deletion	-3.40	Deleterious	159	30	NA	NA	NA	NA	
+10	12,121438957,ACC,.	ENSP00000443112	662	1	GAG [ACC/-] TTC	651	T	.	Deletion	-3.66	Deleterious	158	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000236040	804	-1	CCA [-/TGG] CAT	506	P	PW	Insertion	-9.11	Deleterious	157	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000296388	736	-1	CCA [-/TGG] CAT	506	P	PW	Insertion	-9.11	Deleterious	158	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000380245	697	-1	CCA [-/TGG] CAT	506	P	PW	Insertion	-9.14	Deleterious	160	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000394294	152	-1	CCA [-/TGG] CAT	98	P	PW	Insertion	-9.13	Deleterious	153	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000408056	222	-1	CCA [-/TGG] CAT	114	P	PW	Insertion	-9.17	Deleterious	153	30	NA	NA	NA	NA	
+11	1,43217995,G,GCCA	ENSP00000437795	601	-1	CCA [-/TGG] CAT	371	P	PW	Insertion	-8.74	Deleterious	162	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000314437	631	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.15	Deleterious	137	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000359240	669	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.26	Deleterious	136	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000359243	669	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.26	Deleterious	136	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000402042	199	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.83	Deleterious	101	27	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000407713	240	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-3.24	Deleterious	113	30	NA	NA	NA	NA	
+12	10,102762472,G,GGCG	ENSP00000416972	221	1	CAG [-/GCG] GAT	59	Q	QA	Insertion	-4.09	Deleterious	97	26	NA	NA	NA	NA	
+13	9,117856130,T,G	record not found																rs1513
+14	9,117856135,C,G	record not found																
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/sift_web/test-data/sift_output_summary.tsv	Tue Oct 07 19:19:11 2014 -0400
@@ -0,0 +1,13 @@
+#Type	Total	Found_in_dbSNP	Not_found_in_dbSNP	PROVEAN_neutral	PROVEAN_deleterious	PROVEAN_NA	SIFT_tolerated	SIFT_damaging	SIFT_NA
+Protein_coding	12	3	9	5	6	1	3	2	7
+Single_AA_Change	4	2	2	3	1	0	2	2	0
+Synonymous	2	0	2	2	0	0	1	0	1
+Deletion	3	1	2	0	3	0	0	0	3
+Insertion	2	0	2	0	2	0	0	0	2
+Multiple_AA_Change	0	0	0	0	0	0	0	0	0
+Frameshift	1	0	1	0	0	1	0	0	1
+Nonsense	0	0	0	0	0	0	0	0	0
+Unknown	0	0	0	0	0	0	0	0	0
+Input_error	0	0	0	0	0	0	0	0	0
+Non_protein_coding	2	1	1	0	0	2	0	0	2
+Input_format_error	0	0	0	0	0	0	0	0	0