Feat: add logging and don't watch first page

pdf_oralia/extract.py

@@ -1,3 +1,4 @@
+import logging
 from pathlib import Path
 
 import pdfplumber
@@ -14,11 +15,10 @@ charge_table_settings = {
 def extract_from_pdf(pdf, charge_dest, location_dest):
     """Build charge_dest and location_dest xlsx file from pdf"""
     loc_table = []
-    for page in pdf.pages:
+    for page in pdf.pages[1:]:
+        page_text = page.extract_text()
         situation_loc_line = [
-            l
+            l for l in page_text.split("\n") if "SITUATION DES LOCATAIRES" in l
-            for l in page.extract_text().split("\n")
-            if "SITUATION DES LOCATAIRES" in l
         ]
         if situation_loc_line:
             mois, annee = situation_loc_line[0].split(" ")[-2:]
@@ -27,15 +27,16 @@ def extract_from_pdf(pdf, charge_dest, location_dest):
             else:
                 loc_table = page.extract_table()
 
-        if "HONORAIRES" in page.extract_text():
+        elif "HONORAIRES" in page_text:
-            df_charge = extract_charge(page.extract_table(charge_table_settings))
+            table = page.extract_table(charge_table_settings)
+            df_charge = extract_charge(table)
             df_charge.to_excel(charge_dest, sheet_name="Charges", index=False)
-            print(f"{charge_dest} saved")
+            logging.info(f"{charge_dest} saved")
 
     df_loc = extract_situation_loc(loc_table, mois=mois, annee=annee)
     df_loc = df_loc.assign()
     df_loc.to_excel(location_dest, sheet_name="Location", index=False)
-    print(f"{location_dest} saved")
+    logging.info(f"{location_dest} saved")
 
 
 def extract_save(pdf_file, dest):

pdf_oralia/extract_charge.py

@@ -1,3 +1,5 @@
+import logging
+
 import numpy as np
 import pandas as pd
 
@@ -8,15 +10,6 @@ def extract_charge(table):
         pd.DataFrame(table[1:], columns=table[0])
         .replace("", np.nan)
         .dropna(subset=["Débits"])
-        .astype(
-            {
-                "Débits": "float64",
-                "Crédits": "float64",
-                "Dont T.V.A.": "float64",
-                "Locatif": "float64",
-                "Déductible": "float64",
-            }
-        )
     )
     drop_index = df[
         df["RECAPITULATIF DES OPERATIONS"].str.contains("TOTAUX", case=False)
@@ -30,4 +23,14 @@ def extract_charge(table):
         inplace=True,
     )
 
+    df = df.astype(
+        {
+            "Débits": "float64",
+            "Crédits": "float64",
+            "Dont T.V.A.": "float64",
+            "Locatif": "float64",
+            "Déductible": "float64",
+        }
+    )
+
     return df

pdf_oralia/scripts.py

@@ -1,9 +1,29 @@
+import logging
+from logging.config import dictConfig
 from pathlib import Path
 
 import click
 
 from .extract import extract_save
 
+logging_config = dict(
+    version=1,
+    formatters={"f": {"format": "%(levelname)-8s %(name)-12s %(message)s"}},
+    handlers={
+        "h": {
+            "class": "logging.StreamHandler",
+            "formatter": "f",
+            "level": logging.DEBUG,
+        }
+    },
+    root={
+        "handlers": ["h"],
+        "level": logging.DEBUG,
+    },
+)
+
+dictConfig(logging_config)
+
 
 @click.group()
 def main():
@@ -37,6 +57,7 @@ def all(folder, dest):
 
     pdf_files = [x for x in p.iterdir() if ".pdf" in str(x)]
     for pdf_file in pdf_files:
+        logging.info(f"Found {pdf_file}")
         extract_save(pdf_file, d)