updated version; added smoother handling of data with all records fai…

…ling data checks

Pipfile

@@ -10,5 +10,5 @@ rpy2 = "*"
 [packages]
 pandas = "*"
 numpy = "*"
-vacheck = "*"
+vacheck = ">=0.0.3"
 xlrd = "*"

interva/__version__.py

@@ -1,7 +1,7 @@
 __title__ = "interva"
 __description__ = "Python implementation of the InterVA Algorithm."
 __url__ = "https://github.com/verbal-autopsy-software/interva"
-__version__ = "0.0.6"
+__version__ = "0.0.7"
 __author__ = "Sherry Zhao & Jason Thomas"
 __author_email__ = "[email protected]"
 __license__ = "GPLv3"

interva/interva5.py

@@ -22,6 +22,7 @@
 import datetime
 from pkgutil import get_data
 from io import BytesIO
+import warnings
 
 from interva.data.causetext import CAUSETEXTV5
 from interva.utils import _get_dem_groups
@@ -607,10 +608,16 @@ def run(self) -> None:
                                  "WHOLEPROB"]
             VA_result.drop(nan_indices, axis=0, inplace=True)
         else:
+            # TODO: add get_errors() function (similar to pyinsilicova)
+            warnings.warn(
+                "NO VALID VA RECORDS (datacheck procedure invalidated all "
+                "deaths)!  Check error log for more details."
+            )
             VA_result = None
 
-        dem_group = DataFrame(list_dem_group)
-        self.dem_group = dem_group.set_index("ID")
+        if len(list_dem_group) > 0:
+            dem_group = DataFrame(list_dem_group)
+            self.dem_group = dem_group.set_index("ID")
 
         self.results = {"ID": ID_list,
                         "VA5": VA_result,
@@ -636,21 +643,21 @@ def set_hiv(self, hiv_level: str) -> str:
         hiv_lvl = hiv_level.lower()
         if hiv_lvl in ["h", "l", "v"]:
             self.hiv = hiv_lvl
+            print(f"HIV parameter is {self.hiv}")
         else:
             print(f"The provided HIV level '{hiv_level}' is invalid.")
         return self.hiv
-        print(f"HIV parameter is {self.hiv}")
 
     def set_malaria(self, malaria_level: str) -> str:
         """Set malaria parameter."""
 
         malaria_lvl = malaria_level.lower()
         if malaria_lvl in ["h", "l", "v"]:
             self.malaria = malaria_lvl
+            print(f"Malaria parameter is {self.malaria}")
         else:
             print(f"The provided malaria level '{malaria_level}' is invalid.")
         return self.malaria
-        print(f"Malaria parameter is {self.malaria}")
 
     def get_ids(self) -> Series:
         """Return pandas series of ID column in data."""
@@ -677,6 +684,13 @@ def get_csmf(self, top: int = 10,
         :rtype: pandas.series
         """
 
+        if len(self.results) == 0:
+            print("No results.  Use run() method to assign causes.")
+            return None
+        if self.results["VA5"] is None:
+            print("No results found.  Check error log.  It is likely that "
+                  "all records failed the data consistency checks.")
+            return None
         va = self.results["VA5"]
         set_option("display.max_rows", None)
         set_option("display.max_columns", None)
@@ -738,8 +752,10 @@ def get_csmf(self, top: int = 10,
         dist = None
         for i in range(len(va)):
             if va.iloc[i, 14] is not None:
-                dist = [[0 for _ in range(len(va.iloc[i, 14]))]]
+                # dist = [[0 for _ in range(len(va.iloc[i, 14]))]]
+                dist = array([0] * len(va.iloc[i, 14]))
                 break
+        # what if dist is still None at this point? ex dataset 2?
         undeter = 0
 
         # Pick not simply the top # causes,
@@ -786,10 +802,11 @@ def get_csmf(self, top: int = 10,
                     this_dist[k] = 0
 
                 if va.iloc[i, 14] is not None:
-                    if i == 0:
-                        dist = this_dist
-                    else:
-                        dist = dist + this_dist
+                    # if i == 0:
+                    #     dist = this_dist
+                    # else:
+                    #     dist = dist + this_dist
+                    dist = dist + this_dist
 
         dist = Series(dist)
         dist_cod = None

interva/utils.py

@@ -352,6 +352,7 @@ def _csmf_with_interva_rule(va5: DataFrame) -> Series:
     for i in va.index:
         if va.loc[i, "WHOLEPROB"] is not None:
             dist = zeros(len(va.loc[i, "WHOLEPROB"]))
+            # TODO: this has been changed (to fix bug) in interva5.py
             break
     undetermined = 0