update

graph.py

@@ -1,27 +1,55 @@
 import argparse
-import datetime
+from datetime import datetime
 import pandas as pd
 from pandas import DataFrame
 import plotly.offline as py
 import plotly.graph_objs as go
+from typing import List
 
 
-def plot(df: DataFrame, is_x_time: bool=True, time_format: str="%m/%d/%Y %H:%M:%S", to_jupyter: bool=False):
-    x_values = \
-        df.iloc[:, 0].apply(lambda d: datetime.datetime.strptime(d, time_format)) \
-        if is_x_time else \
-        df.iloc[:, 0]
+def create_graph_datalist(df: DataFrame, path: str="", is_x_time: bool=True, time_format: str="%m/%d/%Y %H:%M:%S", use_elapsed: bool=False, specified_column: str="") -> List[go.Scatter]:
+    x_values = None
+    if is_x_time:
+        if use_elapsed:
+            startdatetime = datetime.strptime(df.iloc[0, 0], time_format)
+            x_values = df.iloc[:, 0].apply(lambda d: (datetime.strptime(d, time_format) - startdatetime).total_seconds())
+        else:
+            x_values = df.iloc[:, 0].apply(lambda d: datetime.strptime(d, time_format))
+    else:
+        x_values = df.iloc[:, 0]
 
-    fig = go.Figure(
-        data=[
+    if specified_column is None or specified_column == "None" or specified_column == "":
+        return [
             go.Scatter(
                 x=x_values,
                 y=df[column_name],
-                name=column_name,
+                name=path + ": " + column_name,
                 opacity=0.50,
             ) for column_name in list(df.columns)[1:]
-        ],
-        layout=go.Layout()
+        ]
+    else:
+        return [
+            go.Scatter(
+                x=x_values,
+                y=df[specified_column],
+                name=path + ": " + specified_column,
+                opacity=0.50,
+            )
+        ]
+
+
+def plot(dfs: List[DataFrame], pathlist: List[str], is_x_time: bool=True, time_format: str="%m/%d/%Y %H:%M:%S", to_jupyter: bool=False, specified_column: str="", use_elapsed: bool=False):
+    data = []
+    for df, path in zip(dfs, pathlist):
+        data += create_graph_datalist(df, path=path, is_x_time=is_x_time, time_format=time_format, use_elapsed=use_elapsed, specified_column=specified_column)
+
+    fig = go.Figure(
+        data=data,
+        layout=go.Layout(
+            title="",
+            xaxis=dict(title="Elapsed [sec]"),
+            yaxis=dict(title="Bytes", exponentformat="SI")
+        )
     )
 
     if not to_jupyter:
@@ -31,18 +59,21 @@ def plot(df: DataFrame, is_x_time: bool=True, time_format: str="%m/%d/%Y %H:%M:%
         py.iplot(fig)
 
 
-def main(file_path: str, encoding: str, is_x_time: bool, time_format: str):
-    df = pd.read_csv(file_path, encoding=encoding)
-    plot(df, is_x_time=is_x_time, time_format=time_format)
+def main(file_path_list: List[str], encoding: str, is_x_time: bool, time_format: str, specified_column: str, use_elapsed: bool):
+    dfs = map(lambda path: pd.read_csv(path, encoding=encoding), file_path_list)
+    # df = pd.read_csv(file_path, encoding=encoding)
+    plot(dfs, file_path_list, is_x_time=is_x_time, time_format=time_format, specified_column=specified_column, use_elapsed=use_elapsed)
 
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser()
-    parser.add_argument("path", help="Path of csv file to read.", type=str)
+    parser.add_argument("path_list", help="Path list of csv files to read.", nargs="*")
     parser.add_argument("--encoding", help="Encoding option of dataframe. Default: utf-16", type=str, default="utf-16")
-    parser.add_argument("--is_x_time", help="Whether to handle the first column as time value or not. Default: True", type=bool, default=True)
+    parser.add_argument("--is_x_time", help="To handle the first column as time value. Default: True", type=lambda s: s.lower() in ["true", "1", "yes"], default="True")
     parser.add_argument("--time_format", help="Time format for the first column. Default: %m/%d/%Y %H:%M:%S", type=str, default="%m/%d/%Y %H:%M:%S")
+    parser.add_argument("--specify_column", help="If column name is specified here, then only plot the column as Y every one file. If not specified, then plot all columns as Y except the first column (X).", type=str, default="None")
+    parser.add_argument("--elapsed", help="To convert the first column to elapsed time. Default: False", type=lambda s: s.lower() in ["true", "1", "yes"], default="False")
 
     args = parser.parse_args()
 
-    main(args.path, args.encoding, args.is_x_time, args.time_format)
+    main(args.path_list, args.encoding, args.is_x_time, args.time_format, args.specify_column, args.elapsed)