aboutsummaryrefslogtreecommitdiff
path: root/wqflask
diff options
context:
space:
mode:
authorLei Yan2014-02-28 00:00:42 +0000
committerLei Yan2014-02-28 00:00:42 +0000
commit07dceea5c550891c00b33f4d665f1de2ec936fea (patch)
treee8b54999a9e9dbf67d6e8e3f0d934aaccdac96f6 /wqflask
parent01a6c3c6c9769f1ab8c30de77441502d403b04b3 (diff)
downloadgenenetwork2-07dceea5c550891c00b33f4d665f1de2ec936fea.tar.gz
Made some changes to Lei's IO code for GN1 and GN2 genofiles
Diffstat (limited to 'wqflask')
-rw-r--r--wqflask/maintenance/dataset/datasampledir/load_genotypes/sample.geno2
-rw-r--r--wqflask/maintenance/dataset/load_genotypes.py221
-rw-r--r--wqflask/maintenance/dataset/utilities.py4
l---------wqflask/maintenance/our_settings.py2
4 files changed, 134 insertions, 95 deletions
diff --git a/wqflask/maintenance/dataset/datasampledir/load_genotypes/sample.geno b/wqflask/maintenance/dataset/datasampledir/load_genotypes/sample.geno
index 0024ffd1..a28d31fc 100644
--- a/wqflask/maintenance/dataset/datasampledir/load_genotypes/sample.geno
+++ b/wqflask/maintenance/dataset/datasampledir/load_genotypes/sample.geno
@@ -9,4 +9,4 @@ Chr Locus cM Mb BXD1 BXD2 BXD5 BXD6 BXD8
2 rs6365999 0.3 4.811062 B B D D D
3 rs6376963 0.895 5.008089 B B D D D
4 rs3677817 1.185 5.176058 B B D D D
-5 rs8236463 2.081 5.579193 B B D D D
+5 rstest8236463 2.081 5.579193 B B D D D
diff --git a/wqflask/maintenance/dataset/load_genotypes.py b/wqflask/maintenance/dataset/load_genotypes.py
index 31aaf1aa..fbf6484f 100644
--- a/wqflask/maintenance/dataset/load_genotypes.py
+++ b/wqflask/maintenance/dataset/load_genotypes.py
@@ -1,31 +1,48 @@
+#Do whatever else is needed with the Marker object
+#Probably create Genofile object as well
+#Make sure rest of code works with params object (though
+#everything in the params object should probably just be the parameters of
+#the Genofile object)
+
+
+from __future__ import absolute_import, print_function, division
+
import sys
import re
+import argparse
import utilities
import datastructure
-def main(argv):
+def main():
+ parser = argparse.ArgumentParser(description='Load Genotypes')
+ parser.add_argument('-c', '--config')
+ opts = parser.parse_args()
+ config = opts.config
# config
- config = utilities.get_config(argv[1])
- print "config:"
+ config = utilities.get_config(config)
+ print("config:")
for item in config.items('config'):
- print "\t%s" % (str(item))
+ print("\t", str(item))
+ parse_genofile(fetch_parameters(config))
+
+def fetch_parameters(config):
# variables
- inbredsetid = config.get('config', 'inbredsetid')
- print "inbredsetid: %s" % inbredsetid
- species = datastructure.get_species(inbredsetid)
- speciesid = species[0]
- print "speciesid: %s" % speciesid
- genofreeze = datastructure.get_genofreeze_byinbredsetid(inbredsetid)
- genofreezeid = genofreeze[0]
- print "genofreezeid: %s" % genofreezeid
- dataid = datastructure.get_nextdataid_genotype()
- print "next data id: %s" % dataid
- cursor, con = utilities.get_cursor()
+ params = {}
+ params['inbredsetid'] = config.get('config', 'inbredsetid')
+ species = datastructure.get_species(params['inbredsetid'])
+ params["speciesid"] = species[0]
+ genofreeze = datastructure.get_genofreeze_byinbredsetid(params['inbredsetid'])
+ params['genofreezeid'] = genofreeze[0]
+ params['dataid'] = datastructure.get_nextdataid_genotype()
+ params['genofile'] = config.get('config', 'genofile')
+ return params
+
+def parse_genofile(params):
# genofile
- genofile = open(config.get('config', 'genofile'), 'r')
+ genofile = open(params['genofile'], 'r')
metadic = {}
- print
+ print()
# parse genofile
for line in genofile:
line = line.strip()
@@ -42,90 +59,112 @@ def main(argv):
continue
if line.lower().startswith("chr"):
#
- print "geno file meta:"
+ print("geno file meta:")
for k, v in metadic.items():
- print "\t%s: %s" % (k, v)
+ print("\t{}: {}".format(k, v))
#
- print "geno file head:\n\t%s" % line
- print
+ print("geno file head:\n\t{}\n".format(line))
strainnames = line.split()[4:]
strains = datastructure.get_strains_bynames(inbredsetid=inbredsetid, strainnames=strainnames, updatestrainxref="yes")
continue
# geno file line
- cells = line.split()
- chr = cells[0]
- locus = cells[1]
- cm = cells[2]
- mb = cells[3]
- values = cells[4:]
- # geno
+ marker = Marker(line)
+ #
+ genoid = check_or_insert_geno(params, marker)
+ if check_genoxref(params):
+ continue
+ insert_genodata(params)
+ insert_genoxref(params)
+ dataid += 1
+ genofile.close()
+
+
+class Marker(object):
+ def __init__(self, line):
+ self.cells = line.split()
+ self.chromosome = cells[0]
+ self.locus = cells[1]
+ self.cm = cells[2]
+ self.mb = cells[3]
+ self.values = cells[4:]
+
+def check_or_insert_geno(params, marker):
+ cursor, con = utilities.get_cursor()
+ sql = """
+ SELECT Geno.`Id`
+ FROM Geno
+ WHERE Geno.`SpeciesId`=%s
+ AND Geno.`Name` like %s
+ """
+ cursor.execute(sql, (speciesid, locus))
+ result = cursor.fetchone()
+ if result:
+ genoid = result[0]
+ print("get geno record: %d" % genoid)
+ else:
sql = """
- SELECT Geno.`Id`
- FROM Geno
- WHERE Geno.`SpeciesId`=%s
- AND Geno.`Name` like %s
+ INSERT INTO Geno
+ SET
+ Geno.`SpeciesId`=%s,
+ Geno.`Name`=%s,
+ Geno.`Marker_Name`=%s,
+ Geno.`Chr`=%s,
+ Geno.`Mb`=%s
"""
- cursor.execute(sql, (speciesid, locus))
- result = cursor.fetchone()
- if result:
- genoid = result[0]
- print "get geno record: %d" % genoid
- else:
- sql = """
- INSERT INTO Geno
- SET
- Geno.`SpeciesId`=%s,
- Geno.`Name`=%s,
- Geno.`Marker_Name`=%s,
- Geno.`Chr`=%s,
- Geno.`Mb`=%s
- """
- cursor.execute(sql, (speciesid, locus, locus, chr, mb))
- rowcount = cursor.rowcount
- genoid = con.insert_id()
- print "INSERT INTO Geno: %d record: %d" % (rowcount, genoid)
- # genodata
- for index, strain in enumerate(strains):
- strainid = strain[0]
- value = utilities.to_db_string(values[index], None)
- if not value:
- continue
- value = config.get('config', "genovalue_" + value)
- try:
- number = int(value)
- except:
- continue
- if not number in [-1, 0, 1]:
- continue
- sql = """
- INSERT INTO GenoData
- SET
- GenoData.`Id`=%s,
- GenoData.`StrainId`=%s,
- GenoData.`value`=%s
- """
- cursor.execute(sql, (dataid, strainid, number))
- # genoxref
+ cursor.execute(sql, (speciesid, locus, locus, chr, mb))
+ rowcount = cursor.rowcount
+ genoid = con.insert_id()
+ print("INSERT INTO Geno: %d record: %d" % (rowcount, genoid))
+ return genoid
+
+def check_GenoXRef():
+ sql = """
+ select GenoXRef.*
+ from GenoXRef
+ where GenoXRef.`GenoFreezeId`=%s
+ AND GenoXRef.`GenoId`=%s
+ """
+ cursor.execute(sql, (genofreezeid, genoid))
+ rowcount = cursor.rowcount
+ return rowcount
+
+def insert_genodata():
+ for index, strain in enumerate(strains):
+ strainid = strain[0]
+ value = utilities.to_db_string(values[index], None)
+ if not value:
+ continue
+ value = config.get('config', "genovalue_" + value)
+ try:
+ number = int(value)
+ except:
+ continue
+ if not number in [-1, 0, 1]:
+ continue
sql = """
- INSERT INTO GenoXRef
+ INSERT INTO GenoData
SET
- GenoXRef.`GenoFreezeId`=%s,
- GenoXRef.`GenoId`=%s,
- GenoXRef.`DataId`=%s,
- GenoXRef.`cM`=%s,
- GenoXRef.`Used_for_mapping`=%s
+ GenoData.`Id`=%s,
+ GenoData.`StrainId`=%s,
+ GenoData.`value`=%s
"""
- cursor.execute(sql, (genofreezeid, genoid, dataid, cm, 'N'))
- rowcount = cursor.rowcount
- print "INSERT INTO GenoXRef: %d record" % (rowcount)
- # for loop next
- dataid += 1
- print
- # release
- genofile.close()
- con.close()
+ cursor.execute(sql, (dataid, strainid, number))
+
+def insert_genoxref():
+ sql = """
+ INSERT INTO GenoXRef
+ SET
+ GenoXRef.`GenoFreezeId`=%s,
+ GenoXRef.`GenoId`=%s,
+ GenoXRef.`DataId`=%s,
+ GenoXRef.`cM`=%s,
+ GenoXRef.`Used_for_mapping`=%s
+ """
+ cursor.execute(sql, (genofreezeid, genoid, dataid, cm, 'N'))
+ rowcount = cursor.rowcount
+ print("INSERT INTO GenoXRef: %d record" % (rowcount))
if __name__ == "__main__":
- print "command line arguments:\n\t%s" % sys.argv
- main(sys.argv)
- print "exit successfully"
+ print("command line arguments:\n\t%s" % sys.argv)
+ main()
+ print("exit successfully")
diff --git a/wqflask/maintenance/dataset/utilities.py b/wqflask/maintenance/dataset/utilities.py
index d389e672..787c9481 100644
--- a/wqflask/maintenance/dataset/utilities.py
+++ b/wqflask/maintenance/dataset/utilities.py
@@ -4,8 +4,8 @@ import ConfigParser
def get_cursor():
host = 'localhost'
- user = 'webqtl'
- passwd = 'webqtl'
+ user = 'gn2'
+ passwd = 'UhHJuiS6gC8hj4a'
db = 'db_webqtl'
con = MySQLdb.Connect(db=db, host=host, user=user, passwd=passwd)
cursor = con.cursor()
diff --git a/wqflask/maintenance/our_settings.py b/wqflask/maintenance/our_settings.py
index 14efe407..b14de960 120000
--- a/wqflask/maintenance/our_settings.py
+++ b/wqflask/maintenance/our_settings.py
@@ -1 +1 @@
-../../../zach_settings.py \ No newline at end of file
+../../../lei_settings.py \ No newline at end of file