--- a
+++ b/prescreen/evaluation/fetch_radio.py
@@ -0,0 +1,71 @@
+"""
+fetches radiology reports from validation cohort
+
+"""
+import pandas as pd
+
+
+def get_frames():
+    base_path = 'data/cohorte_validation'
+
+    # regular expression to find radiology reports
+    pattern = "recist"
+
+
+    ## DITEP OK
+    path = base_path + '/ditep_inclus.csv'
+
+    ditep_inclus = pd.read_csv(path, sep=';', encoding='utf-8')\
+        .drop('Unnamed: 0', axis=1)
+
+    mask_scanner = ditep_inclus['CR'].str.contains(pattern, case=False)
+
+    ditep_inclus = ditep_inclus[mask_scanner]
+    ditep_inclus = ditep_inclus.groupby('nip', as_index=False).agg('first')
+
+
+
+    ## DITEP screenfail
+    path = base_path + '/ditep_sf.csv'
+    ditep_sf = pd.read_csv(path, sep=';', encoding='utf-8')\
+        .drop('Unnamed: 0', axis=1)
+
+    mask_scanner = ditep_sf['CR'].str.contains(pattern, case=False)
+
+    ditep_sf = ditep_sf[mask_scanner]
+    ditep_sf = ditep_sf.groupby('nip', as_index=False).agg('first')
+
+
+    ## poumons inclus
+    path = base_path + '/poumons_inclusion.csv'
+
+    poumons = pd.read_csv(path, sep=';', encoding='utf-8')\
+        .drop('Unnamed: 0', axis=1)
+
+    mask_scanner = poumons['CR'].str.contains(pattern, case=False)
+
+    poumons = poumons[mask_scanner]
+    poumons = poumons.groupby('nip', as_index=False).agg('first')
+
+    # poumons SF
+    path = base_path + '/poumons_sf.csv'
+
+    poumons_sf = pd.read_csv(path, sep=';', encoding='utf-8')\
+        .drop('Unnamed: 0', axis=1)
+
+    mask_scanner = poumons_sf['CR'].str.contains(pattern, case=False)
+
+    poumons_sf = poumons_sf[mask_scanner]
+    poumons_sf = poumons_sf.groupby('nip', as_index=False).agg('first')
+
+
+
+    return ditep_inclus, ditep_sf, poumons, poumons_sf
+
+
+
+
+
+
+
+