check for arxiv before try to get local title

WissamAntoun · Jun 30, 2020 · d9fd308 · d9fd308
1 parent ed3a534
commit d9fd308
Show file tree

Hide file tree

Showing 3 changed files with 34 additions and 30 deletions.
diff --git a/.vscode/settings.json b/.vscode/settings.json
@@ -0,0 +1,3 @@
+{
+    "python.pythonPath": "C:\\Users\\WISSAM-PC\\.conda\\envs\\thesis\\python.exe"
+}
diff --git a/rename_pdf_files_from_content_title_or_arxiv.py b/rename_pdf_files_from_content_title_or_arxiv.py
@@ -4,7 +4,7 @@
 import re
 import glob
 import pdftitle
-from tqdm import tqdm_notebook as tqdm
+from tqdm import tqdm
 import arxiv
 import random
 #%%
@@ -25,7 +25,7 @@ def get_valid_filename(s):
     return re.sub(r'(?u)[^-\w.]', '', s)
 
 #%%
-all_files_names = glob.glob("C:\\Users\\WISSAM-PC\\Downloads\\Documents - copy\\*.pdf")
+all_files_names = glob.glob("C:\\Users\\WISSAM-PC\\Downloads\\Documents\\*.pdf")
 #%%
 # research_files = []
 # for f in tqdm(all_files_names):
@@ -39,35 +39,36 @@ def get_valid_filename(s):
 # %%
 research_files = []
 for f in tqdm(all_files_names):
-    basename = os.path.basename(f)
-    if "__OLDNAME__" in basename:
+    if "__OLDNAME__" in f or '__XX__' in f:
         continue
+    basename = os.path.basename(f)
     dirname = os.path.dirname(f)
-    try:
-        title = pdftitle.get_title_from_file(f)
-        if len(title) < 3:
-            continue
-        title = re.sub(r'[\\/:"*?<>|]+'," ",title)
-        title = re.sub(r'\s'," ",title)
-        title = get_valid_filename(title)
-        new_basename = title + '__OLDNAME__' + basename
-        new_path = os.path.join(dirname, new_basename)
-    except:
-        title = None
+
+    if len(re.findall(r'\b\d{4}\.\d{5}',basename)) > 0 :
+        article_id = basename.strip('.pdf')
+        entry = arxiv.query(id_list=[article_id])
+        if len(entry) > 0:
+            title = entry[0]['title'].replace(":","")
+            title = re.sub(r'\s'," ",title)
+            title = get_valid_filename(title)
+            new_basename = title + '__OLDNAME__' + basename
+            new_path = os.path.join(dirname, new_basename)
 
-    if title==None:
-        if len(re.findall(r'\b\d{4}\.\d{5}',basename)) > 0 :
-            article_id = basename.strip('.pdf')
-            entry = arxiv.query(id_list=[article_id])
-            if len(entry) > 0:
-                title = entry[0]['title'].replace(":","")
-                title = re.sub(r'\s'," ",title)
-                title = get_valid_filename(title)
-                new_basename = title + '__OLDNAME__' + basename
-                new_path = os.path.join(dirname, new_basename)
-        else:
+    else:
+        try:
+            title = pdftitle.get_title_from_file(f)
+            if len(title) < 3:
+                continue
+            title = re.sub(r'[\\/:"*?<>|]+'," ",title)
+            title = re.sub(r'\s'," ",title)
+            title = get_valid_filename(title)
+            new_basename = title + '__OLDNAME__' + basename
+            new_path = os.path.join(dirname, new_basename)
+        except:
+            title = None
+            os.rename(f,f[:-4]+'__XX__'+'.pdf')
             continue
-    
+
     try:
         print(new_path)
         os.rename(f, new_path)
@@ -76,9 +77,8 @@ def get_valid_filename(s):
         dirname = os.path.dirname(new_path)
         os.rename(f,os.path.join(dirname, new_basename[:-4]+str(random.randint(0,999))+'.pdf'))
     except:
+        os.rename(f,f[:-4]+'__XX__'+'.pdf')
         print('skipping')
 
 
-
-
- # %%
+# %%
diff --git a/run_pdf_renamer.bat b/run_pdf_renamer.bat
@@ -0,0 +1 @@
+"C:\Users\WISSAM-PC\.conda\envs\thesis\python.exe" rename_pdf_files_from_content_title_or_arxiv.py
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		"C:\Users\WISSAM-PC\.conda\envs\thesis\python.exe" rename_pdf_files_from_content_title_or_arxiv.py