Add default confidence param to sssom file while creating ptable (#…

…388)

* Add default `confidence` of 0.95 to sssom file

* added docstring

* updated docstring

* refactored code

* correction

* fixed

* switched `default_confidence` from flag to float. default confidence multiplied to existing if not null

* reformatted

* typos fixed

* removed commented and redundant code

* removed unnecessary imports

* Improved logic

* typo

* fixed casting

* changes to address None float

* Nonee => np.Nan

* changed type to float range

* added space

* removed space

* added space in comment to test something

src/sssom/cli.py

@@ -6,7 +6,7 @@
 - When you run ``python3 -m sssom`` python will execute``__main__.py`` as a script. That means there won't be any
   ``sssom.__main__`` in ``sys.modules``.
 - When you import __main__ it will get executed again (as a module) because
-  there's no ``sssom.__main__`` in ``sys.modules``.
+  there's no ``sssom.__main__`` in ``sys.modules`` .
 
 .. seealso:: https://click.palletsprojects.com/en/8.0.x/setuptools/
 """
@@ -49,7 +49,6 @@
     SSSOM_EXPORT_FORMATS,
     SSSOM_READ_FORMATS,
     MappingSetDataFrame,
-    collapse,
     compare_dataframes,
     dataframe_to_ptable,
     filter_redundant_rows,
@@ -253,14 +252,18 @@ def split(input: str, output_directory: str):
 @input_argument
 @output_option
 @click.option("-W", "--inverse-factor", help="Inverse factor.")
-def ptable(input, output: TextIO, inverse_factor):
+@click.option(
+    "--default-confidence",
+    type=click.FloatRange(0, 1),
+    help="Default confidence to be assigned if absent.",
+)
+def ptable(input, output: TextIO, inverse_factor: float, default_confidence: float):
     """Convert an SSSOM file to a ptable for kboom/`boomer <https://github.com/INCATools/boomer>`_."""
     # TODO should maybe move to boomer (but for now it can live here, so cjm can tweak
     msdf = parse_sssom_table(input)
-    # df = parse(input)
-    df = collapse(msdf.df)
-    # , priors=list(priors)
-    rows = dataframe_to_ptable(df, inverse_factor=inverse_factor)
+    rows = dataframe_to_ptable(
+        msdf.df, inverse_factor=inverse_factor, default_confidence=default_confidence
+    )
     for row in rows:
         print(*row, sep="\t", file=output)
 

src/sssom/util.py

@@ -504,17 +504,41 @@ def compare_dataframes(df1: pd.DataFrame, df2: pd.DataFrame) -> MappingSetDiff:
     return d
 
 
-def dataframe_to_ptable(df: pd.DataFrame, *, inverse_factor: float = None):
+def add_default_confidence(df: pd.DataFrame, confidence: float = np.NAN) -> pd.DataFrame:
+    """Add `confidence` column to DataFrame if absent and initializes to 0.95.
+
+    If `confidence` column already exists, only fill in the None ones by 0.95.
+
+    :param df: DataFrame whose `confidence` column needs to be filled.
+    :return: DataFrame with a complete `confidence` column.
+    """
+    if CONFIDENCE in df.columns:
+        df[CONFIDENCE] = df[CONFIDENCE].apply(lambda x: confidence * x if x is not None else x)
+        df[CONFIDENCE].fillna(float(confidence), inplace=True)
+    else:
+        df[CONFIDENCE] = float(confidence)
+
+    return df
+
+
+def dataframe_to_ptable(
+    df: pd.DataFrame, *, inverse_factor: float = None, default_confidence: float = None
+):
     """Export a KBOOM table.
 
     :param df: Pandas DataFrame
     :param inverse_factor: Multiplier to (1 - confidence), defaults to 0.5
+    :param default_confidence: Default confidence to be assigned if absent.
     :raises ValueError: Predicate value error
     :raises ValueError: Predicate type value error
     :return: List of rows
     """
     if not inverse_factor:
         inverse_factor = 0.5
+
+    if default_confidence:
+        df = add_default_confidence(df, default_confidence)
+
     df = collapse(df)
     rows = []
     for _, row in df.iterrows():