Create test geno iterator

2 files changed, 32 insertions, 2 deletions

diff --git a/wqflask/wqflask/my_pylmm/pyLMM/runlmm.py b/wqflask/wqflask/my_pylmm/pyLMM/runlmm.py
index ef0bdd7e..5a4bd268 100644
--- a/wqflask/wqflask/my_pylmm/pyLMM/runlmm.py
+++ b/wqflask/wqflask/my_pylmm/pyLMM/runlmm.py
@@ -99,11 +99,16 @@ if options.pheno:
     y = tsvreader.pheno(options.pheno)
     print y.shape
 
-if options.geno:
+if options.geno and cmd != 'iterator':
     g = tsvreader.geno(options.geno)
     print g.shape
 
-if cmd == 'redis_new':
+if cmd == 'iterator':
+    print "ITERATE over SNPs"
+    def pretty(snpid,values):
+        print snpid,values
+    print tsvreader.geno_iter(options.geno,pretty)
+elif cmd == 'redis_new':
     # The main difference between redis_new and redis is that missing
     # phenotypes are handled by the first
     if options.remove_missing_phenotypes:
diff --git a/wqflask/wqflask/my_pylmm/pyLMM/tsvreader.py b/wqflask/wqflask/my_pylmm/pyLMM/tsvreader.py
index b4027fa3..7fe75e3f 100644
--- a/wqflask/wqflask/my_pylmm/pyLMM/tsvreader.py
+++ b/wqflask/wqflask/my_pylmm/pyLMM/tsvreader.py
@@ -74,3 +74,28 @@ def geno(fn):
     G = np.array(G1)
     return G
 
+def geno(fn):
+    G1 = []
+    def append(id,values):
+        G1.append(values) # <--- slow
+    geno_iter(fn,append)
+    G = np.array(G1)
+    return G
+
+def geno_iter(fn,func):
+    hab_mapper = {'A':0,'H':1,'B':2,'-':3}
+    pylmm_mapper = [ 0.0, 0.5, 1.0, float('nan') ]
+
+    print fn
+    with open(fn,'r') as tsvin:
+        assert(tsvin.readline().strip() == "# Genotype format version 1.0")
+        tsvin.readline()
+        tsvin.readline()
+        tsvin.readline()
+        tsvin.readline()
+        tsv = csv.reader(tsvin, delimiter='\t')
+        for row in tsv:
+            id = row[0]
+            gs = list(row[1])
+            gs2 = [pylmm_mapper[hab_mapper[g]] for g in gs]
+            func(id,gs2)