From f2bcf6241a8b9a2347764a09d22bfbbfd0feb6ef Mon Sep 17 00:00:00 2001
From: Bertrand Benjamin <benjamin.bertrand@opytex.org>
Date: Thu, 5 Oct 2023 15:10:39 +0200
Subject: [PATCH 1/3] Fix: rebuild join_excel

---
 pdf_oralia/join.py    | 22 ++++++++++++++++++++++
 pdf_oralia/scripts.py |  6 +++++-
 2 files changed, 27 insertions(+), 1 deletion(-)
 create mode 100644 pdf_oralia/join.py

diff --git a/pdf_oralia/join.py b/pdf_oralia/join.py
new file mode 100644
index 0000000..5cf2688
--- /dev/null
+++ b/pdf_oralia/join.py
@@ -0,0 +1,22 @@
+import glob
+
+import pandas as pd
+
+
+def join_excel(src, dest, file_pattern):
+    """Join every excel file in arc respecting file_pattern into on unique file in dist"""
+    filenames = list_files(src, file_pattern)
+    dfs = extract_dfs(filenames)
+    joined_df = pd.concat(dfs)
+    joined_df.to_excel(dest)
+
+
+def list_files(src, file_glob):
+    return list(glob.iglob(f"{src}/{file_glob}"))
+
+
+def extract_dfs(filenames):
+    dfs = []
+    for filename in filenames:
+        dfs.append(pd.read_excel(filename))
+    return dfs
diff --git a/pdf_oralia/scripts.py b/pdf_oralia/scripts.py
index cd486bd..d9e29c1 100644
--- a/pdf_oralia/scripts.py
+++ b/pdf_oralia/scripts.py
@@ -5,6 +5,7 @@ from pathlib import Path
 import click
 
 from .extract import extract_save
+from .join import join_excel
 
 logging_config = dict(
     version=1,
@@ -65,4 +66,7 @@ def all(src, dest):
 @click.option("--src", help="Tous les fichiers dans src", default="./")
 @click.option("--dest", help="Où mettre les fichiers produits", default="")
 def join(src, dest):
-    join_excel(src, dest, df_names=["charge", "locataire"])
+    dest_charge = f"{dest}/charge.xlsx"
+    join_excel(src, dest_charge, "*_charge.xlsx")
+    dest_locataire = f"{dest}/locataire.xlsx"
+    join_excel(src, dest_locataire, "*_locataire.xlsx")
-- 
2.45.2


From 132e37267b168d1bdf48009c475fc4116b50af16 Mon Sep 17 00:00:00 2001
From: Bertrand Benjamin <benjamin.bertrand@opytex.org>
Date: Thu, 5 Oct 2023 15:19:16 +0200
Subject: [PATCH 2/3] Feat: logging and option about overwritting

---
 pdf_oralia/scripts.py | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

diff --git a/pdf_oralia/scripts.py b/pdf_oralia/scripts.py
index d9e29c1..ba4be5f 100644
--- a/pdf_oralia/scripts.py
+++ b/pdf_oralia/scripts.py
@@ -65,8 +65,29 @@ def all(src, dest):
 @main.command()
 @click.option("--src", help="Tous les fichiers dans src", default="./")
 @click.option("--dest", help="Où mettre les fichiers produits", default="")
-def join(src, dest):
+@click.option(
+    "--force",
+    help="Ecraser si le ficher destination existe.",
+    default=False,
+    is_flag=True,
+)
+def join(src, dest, force):
+    """Join tous les fichiers excel charge (resp locataire) de src dans un seul fichier charge.xlsx dans dist.
+
+    Exemple:
+
+        pdf-oralia join --src <dossier_source> --dest <dossier_destination>
+
+
+    """
     dest_charge = f"{dest}/charge.xlsx"
-    join_excel(src, dest_charge, "*_charge.xlsx")
+    if not force and Path(dest_charge).exists():
+        raise ValueError(f"The file {dest_charge} already exists")
     dest_locataire = f"{dest}/locataire.xlsx"
+    if not force and Path(dest_locataire).exists():
+        raise ValueError(f"The file {dest_locataire} already exists")
+
+    join_excel(src, dest_charge, "*_charge.xlsx")
+    logging.info(f"Les données charges ont été concaténées dans {dest_charge}")
     join_excel(src, dest_locataire, "*_locataire.xlsx")
+    logging.info(f"Les données locataires ont été concaténées dans {dest_locataire}")
-- 
2.45.2


From 48e75358ac0a161d49bfdcebe16b45f6c75c8244 Mon Sep 17 00:00:00 2001
From: Bertrand Benjamin <benjamin.bertrand@opytex.org>
Date: Thu, 5 Oct 2023 15:22:14 +0200
Subject: [PATCH 3/3] Fix: remove index in excel outputs

---
 pdf_oralia/join.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pdf_oralia/join.py b/pdf_oralia/join.py
index 5cf2688..37c5aea 100644
--- a/pdf_oralia/join.py
+++ b/pdf_oralia/join.py
@@ -8,7 +8,7 @@ def join_excel(src, dest, file_pattern):
     filenames = list_files(src, file_pattern)
     dfs = extract_dfs(filenames)
     joined_df = pd.concat(dfs)
-    joined_df.to_excel(dest)
+    joined_df.to_excel(dest, index=False)
 
 
 def list_files(src, file_glob):
-- 
2.45.2