pdf-auto-outline 0.1.4__py3-none-any.whl → 0.1.6__py3-none-any.whl

pdf_auto_outline/__init__.py

@@ -1 +1 @@
-__version__ = '0.1.4'
+__version__ = '0.1.6'

pdf_auto_outline/main.py

@@ -1,10 +1,10 @@
 import pymupdf.layout
-from pymupdf import Point
 from time import perf_counter
 from multiprocessing import Pool
 import os
 import subprocess
 import argparse
+import tempfile
 
 SIOYEK = None
 
@@ -148,7 +148,12 @@ def align_toc_lvls(toc_entries: list) -> list:
 
     return toc_entries
 
-def generate_txtfile(toc_entries, txtfile='outline.txt') -> str:
+def get_tmpfile():
+    return tempfile.NamedTemporaryFile(
+            mode='w+', encoding='utf-8', delete=False, suffix='.txt'
+            )
+
+def generate_txtfile(toc_entries, txtfile=get_tmpfile()):
     import textwrap
     txt = textwrap.dedent("""\
     ============================================================
@@ -159,41 +164,46 @@ def generate_txtfile(toc_entries, txtfile='outline.txt') -> str:
     ============================================================
 
     """)
-    if len(toc_entries[0]) > 3:
+    if not toc_entries:
+        pass
+    elif len(toc_entries[0]) > 3:
         txt += '\n'.join(f"{' '*4 * (i[0] - 1)}{i[1]}  |  {i[2]}  |  {i[3]}" 
                         for i in toc_entries)
     else:
         txt += '\n'.join(f"{' '*4 * (i[0] - 1)}{i[1]}  |  {i[2]}"
                         for i in toc_entries)
 
-    with open(txtfile, 'w', encoding='utf-8') as f:
-        f.write(txt)
+    txtfile.write(txt)
+    txtfile.flush()
+    txtfile.seek(0)
 
     return txtfile
 
 
-def parse_txtfile(txtfile='outline.txt', tablevel=2) -> list:
+
+def parse_txtfile(f, tablevel=2) -> list:
     toc_entries = []
-    with open(txtfile) as f:
-        if (c := f.read(1)) == 'C':
-            log('Outline not written')
-            exit()
-        elif c == '=':
-            lines = f.readlines()[7:]
-        else: lines = f.read()
-
-        for i in lines:
-            i = i.replace('\t', '    '*tablevel)
-            lvl = (len(i) - len(i.lstrip())) // 4 + 1
-            a = i.lstrip().split('  |  ')
-            if len(a) < 3:
-                toc_entries.append(
-                        [lvl, a[0], int(a[1])] 
-                )
-            else:
-                toc_entries.append(
-                        [lvl, a[0], int(a[1]), eval(a[2])]
-                )
+    if (c := f.read(1)) == 'C':
+        log('Outline not written')
+        exit()
+    elif c == '=':
+        lines = f.readlines()[7:]
+    else: lines = f.read()
+
+    for i in lines:
+        i = i.replace('\t', '    '*tablevel)
+        lvl = (len(i) - len(i.lstrip())) // 4 + 1
+        a = i.lstrip().split('  |  ')
+        if len(a) < 3:
+            toc_entries.append(
+                    [lvl, a[0], int(a[1])] 
+            )
+        else:
+            toc_entries.append(
+                    [lvl, a[0], int(a[1]), eval(a[2])]
+            )
+    
+    f.close()
 
     return toc_entries
 
@@ -208,18 +218,15 @@ def embed_toc(pdfpath, toc_entries, newfile=''):
         log(f"toc saved to '{pdfpath}'")
 
 def get_toc_custom(doc) -> list:
-    toc_entries = [[*i[:3], i[3].get('to')] for i in doc.get_toc(False)]
+    toc_entries = [[*i[:3], i[3].get('to')[1]] for i in doc.get_toc(False)]
     return toc_entries
 
-def edit_txtfile(txtfile='outline.txt'):
-    # editor = os.environ.get('EDITOR', 'notepad' if os.name == 'nt' else 'vi')
-    # editor = os.environ.get('EDITOR', 'start' if os.name == 'nt' else 'xdg-open')
-    name = os.name
-    if name == 'nt':
-        subprocess.run(['start', '/WAIT', txtfile], shell=True)
+def edit_txtfile(f):
+    if os.name == 'nt':
+        subprocess.run(['start', '/WAIT', f.name], shell=True)
     else: # name == 'posix':
         editor = os.environ.get('EDITOR', 'vi')  
-        subprocess.run([editor, txtfile])
+        subprocess.run([editor, f.name])
 
 def main():
     parser = argparse.ArgumentParser(prog='pdfao')
@@ -232,19 +239,19 @@ def main():
     parser.add_argument('-i', '--infile', type=str, metavar='<file>', help='write toc from file to pdf')
     parser.add_argument('-t', '--tablevel', type=int, metavar='<n>', help='tab = n toc nesting levels (default 2)', default=2)
     parser.add_argument('--sioyek', type=str, metavar='<path>', help='for users of the Sioyek pdf viewer')
-    parser.add_argument('--version', action='version', version='%(prog)s 0.1.4')
+    parser.add_argument('--version', action='version', version='%(prog)s 0.1.6')
 
     args = parser.parse_args()
 
+    if args.out:
+        args.out = os.path.join(
+                os.path.dirname(args.filename),
+                args.out)
+
     if args.sioyek:
         from sioyek.sioyek import Sioyek
         global SIOYEK
         SIOYEK = Sioyek(args.sioyek)
-        if args.out:
-            args.out = os.path.join(
-                    os.path.dirname(args.filename),
-                    args.out
-                )
         # local_db = args.sioyek[1]
         # shared_db = args.sioyek[2]
         # pdf_path = args.sioyek[3]
@@ -253,14 +260,15 @@ def main():
     if args.edit or args.superedit:
         doc = pymupdf.Document(args.filename)
         if args.superedit:
-            generate_txtfile(doc.get_toc(False))
+            f = generate_txtfile(doc.get_toc(False))
         else:
-            generate_txtfile(get_toc_custom(doc))
-        edit_txtfile()
-        toc_entries = parse_txtfile(tablevel=args.tablevel)
+            f = generate_txtfile(get_toc_custom(doc))
+        edit_txtfile(f)
+        toc_entries = parse_txtfile(f, args.tablevel)
         embed_toc(args.filename, toc_entries, args.out)
+        os.remove(f.name)
     elif args.infile:
-        toc_entries = parse_txtfile(args.infile, args.tablevel)
+        toc_entries = parse_txtfile(open(args.infile, encoding='utf-8'), args.tablevel)
         embed_toc(args.filename, toc_entries, args.out)
     else: # generate toc
         start = perf_counter()
@@ -272,10 +280,11 @@ def main():
         if args.straight:
             embed_toc(args.filename, toc_entries, args.out)
         else:
-            generate_txtfile(toc_entries)
-            edit_txtfile()
-            toc_entries = parse_txtfile(tablevel=args.tablevel)
+            f = generate_txtfile(toc_entries)
+            edit_txtfile(f)
+            toc_entries = parse_txtfile(f, args.tablevel)
             embed_toc(args.filename, toc_entries, args.out)
+            os.remove(f.name)
 
     # if args.sioyek and not args.out:
     #     to_hash = get_md5_hash(args.filename)

pdf_auto_outline/test.py

@@ -0,0 +1,4 @@
+txt = "aa\nbb\ncc"
+
+for i in txt:
+    print(i)

{pdf_auto_outline-0.1.4.dist-info → pdf_auto_outline-0.1.6.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: pdf-auto-outline
-Version: 0.1.4
+Version: 0.1.6
 Summary: Automatically generate and edit PDF table of contents / outline
 Author: Rossikos
 Author-email: Rossikos <216631970+rossikos@users.noreply.github.com>
@@ -62,6 +62,28 @@ Edit exiting pdf toc:
 A save toc to new pdf from file:
 `pdfao paper.pdf -o new.pdf -i outline.txt`
 
+### Editing
+
+The edit command opens the TOC in the OS default editor (result of 'start' command on Windows and 'EDITOR' environment variable on MacOS and Linux). The file schema is something like this:
+
+```
+Title 1  |  1
+    Title 2  |  2  |  *
+                 ^^^^^^
+				 optional		
+```
+The essential parts of each line are:
+- Indentation - 4 space characters per nesting level (or use tabs with the -t flag).
+- Title text
+- Delimiter - '  |  ' (vertical bar with 2 spaces padding on each side)
+- Page number
+
+The optional part can be one of:
+```
+  |  None                  	same as not including it   
+  |  241.2					y-ordinate 
+  |  {<dictionary>}			dictionary with more attributes for the ToC entry
+```
 
 ## For Sioyek Users
 
@@ -69,8 +91,10 @@ Example commands; add to `prefs_user.config`.
 
 ```
 new_command _gen_toc pdfao "%{file_path}" --sioyek path/to/sioyek -mp 4
-new_command _edit_toc pdfao %{file_path}" --sioyek path/to/sioyek -e
+new_command _edit_toc pdfao "%{file_path}"
 ```
 
+The sioyek library and flag are optional; they allow logging to the status bar. This is more useful for ToC generation where you may want a progress bar. 
+
 If you don't wish to install from PyPI, download source and use `python3 -m path/to/src/pdf_auto_outline` in place of `pdfao`.
 

pdf_auto_outline-0.1.6.dist-info/RECORD

@@ -0,0 +1,8 @@
+pdf_auto_outline/__init__.py,sha256=gW5NUxwGdPsiQjn0cOuuQT11pfthByI5DITDg_HMhLQ,22
+pdf_auto_outline/__main__.py,sha256=mRKsAFeG5R17vTYubIKregAve4vnKc-nk7jY3tcK4wI,78
+pdf_auto_outline/main.py,sha256=AWbXZNVw5e34v005qtoCGyTzNbT3T1mOo6xeLl-Q82k,9883
+pdf_auto_outline/test.py,sha256=2wIjM8xqnDjHiokFF0WUNw0o_oTGW91V-MifHoLw4Rw,47
+pdf_auto_outline-0.1.6.dist-info/WHEEL,sha256=5w2T7AS2mz1-rW9CNagNYWRCaB0iQqBMYLwKdlgiR4Q,78
+pdf_auto_outline-0.1.6.dist-info/entry_points.txt,sha256=HBvhmxJs8hHqbbpJmVTbBH3xy19Hk655O_ySwFC_53w,100
+pdf_auto_outline-0.1.6.dist-info/METADATA,sha256=Rw9x1kQxghKSGVu31qbnLASX87QMjAeUzg2JMAecupU,3287
+pdf_auto_outline-0.1.6.dist-info/RECORD,,

pdf_auto_outline-0.1.4.dist-info/RECORD

@@ -1,7 +0,0 @@
-pdf_auto_outline/__init__.py,sha256=aBEbDvx4LMg8A1TJJR6dEHu8rQODVin528hLS_EDvuA,22
-pdf_auto_outline/__main__.py,sha256=mRKsAFeG5R17vTYubIKregAve4vnKc-nk7jY3tcK4wI,78
-pdf_auto_outline/main.py,sha256=s9rVJJjzqavhAfCAmCOadcOn_C_hbaTmNGSK-WX79JY,9978
-pdf_auto_outline-0.1.4.dist-info/WHEEL,sha256=5w2T7AS2mz1-rW9CNagNYWRCaB0iQqBMYLwKdlgiR4Q,78
-pdf_auto_outline-0.1.4.dist-info/entry_points.txt,sha256=HBvhmxJs8hHqbbpJmVTbBH3xy19Hk655O_ySwFC_53w,100
-pdf_auto_outline-0.1.4.dist-info/METADATA,sha256=ALWzcv1m6sCGDJ6dJyLu9i0RO8IIgNNh086z_Wx873E,2449
-pdf_auto_outline-0.1.4.dist-info/RECORD,,