consolidate manual-collate functionality into collate

dlibrary/dlibrary.py

@@ -214,11 +214,52 @@ def self_and_parents(path):
     return [path] + list(path.parents)
 
 def collate(args):
-    con = sqlite3.connect(args.destdir / 'meta.db')
-    cur = con.cursor()
-
     extraction_dir = args.destdir / 'extract'
-    hint_map = {self_and_parents(Path(relpath(hint, extraction_dir)))[-2].name: hint for hint in args.hints}
+
+    def extracted_path_work_id(path):
+        trail = self_and_parents(Path(relpath(path, extraction_dir)))
+        if len(trail) < 2:
+            return None
+        result = trail[-2].name
+        if result == '..':
+            return None
+        return result
+
+    (raw_groups, raw_exclusions) = parse_expressions(args.expression)
+
+    specified_works = set()
+    works_groups = {}
+    for group in raw_groups:
+        if len(group) == 0:
+            continue
+        work_id = extracted_path_work_id(group[0])
+        if not work_id:
+            print(f'Group {group} contains paths outside an extracted work!')
+            exit(1)
+        if not all(extracted_path_work_id(item) == work_id for item in group[1:]):
+            print(f'Group {group} contains paths from multiple works!')
+            exit(1)
+        specified_works.add(work_id)
+        if work_id not in works_groups:
+            works_groups[work_id] = []
+        normalized_paths = [normalize_to(item, args.destdir) for item in group]
+        if not all(path.exists() for path in normalized_paths):
+            print(f'Group {group} contains nonexistent paths!')
+            exit(1)
+        works_groups[work_id].append(normalized_paths)
+
+    exclusions = []
+    for exclusion in raw_exclusions:
+        work_id = extracted_path_work_id(exclusion)
+        if not work_id:
+            print(f'Excluded path {exclusion} does not belong to an extracted work!')
+            exit(1)
+        specified_works.add(work_id)
+        normalized_path = normalize_to(exclusion, args.destdir)
+        if not normalized_path.exists():
+            print(f'Excluded path {exclusion} does not exist!')
+            exit(1)
+        exclusions.append(normalized_path)
 
     collation_staging_area = args.destdir / 'site' / 'images-staging'
     collation_staging_area.mkdir(parents=True)
@@ -226,21 +267,41 @@ def collate(args):
     collation_area = args.destdir / 'site' / 'images'
     collation_area.mkdir(parents=True, exist_ok=True)
 
+    con = sqlite3.connect(args.destdir / 'meta.db')
+    cur = con.cursor()
+
     for work_path in extraction_dir.iterdir():
         work_id = work_path.name
 
+        if args.only_specified_works and work_id not in specified_works:
+            continue
+
         work_collation_dir = collation_area / work_id
         if work_collation_dir.exists():
+            if work_id not in specified_works:
                 continue
+            if len(list(work_collation_dir.iterdir())) > 0:
+                print(f'Collation directory for work {work_id} already exists!')
+                break
+            else:
+                work_collation_dir.rmdir()
 
         virtual = cur.execute("SELECT virtual FROM works WHERE id = ?", (work_id,)).fetchone()
         if virtual == (1,):
+            if work_id in specified_works:
+                print(f'Work {work_id} is virtual!')
+                break
             continue
 
         work_staging_dir = collation_staging_area / work_id
 
-        collator = Collator(work_staging_dir, [], args)
+        collator = Collator(work_staging_dir, exclusions, args)
-        collation_result = collator.collate_from_paths([hint_map.get(work_id, work_path)])
+        for group in works_groups.get(work_id, [[work_path]]):
+            collation_result = collator.collate_from_paths([item for item in group if item not in exclusions])
+            if not collation_result:
+                print(f'Unable to deduce file structure for {work_id} subgroup {[str(path) for path in group]}')
+                break
+
         if collation_result and collator.index > 0:
             print(f'Collated {collator.index} pages for {work_id}')
             work_staging_dir.rename(work_collation_dir)
@@ -253,7 +314,7 @@ def collate(args):
             if not collation_result:
                 print(f'Unable to deduce file structure for {work_id}, skipping')
             elif collator.index == 0:
-                print(f'{work_id} contains no files? skipping')
+                print(f'No files found for {work_id}, skipping')
 
     collation_staging_area.rmdir()
     con.close()
@@ -688,9 +749,6 @@ def superior_or_equal(a, b):
     return len(a) >= len(b) and all(a[i] >= b[i] for i in range(len(b)))
 
 
-def self_and_parents(path):
-    return [path] + list(path.parents)
-
 def parse_expressions(tokens):
     groups = []
     exclusions = []
@@ -727,61 +785,6 @@ def parse_group(tokens):
 def normalize_to(path, ref):
     return ref / Path(relpath(path, ref))
 
-def manual_collate(args):
-    (raw_groups, raw_exclusions) = parse_expressions(args.expression)
-
-    extraction_dir = args.destdir / 'extract'
-
-    sample_path = next(path for group in (raw_groups + [raw_exclusions]) for path in group)
-    work_id = self_and_parents(Path(relpath(sample_path, extraction_dir)))[-2].name
-
-    exclusions = [normalize_to(item, args.destdir) for item in raw_exclusions]
-
-    if raw_groups:
-        groups = [[normalize_to(item, args.destdir) for item in group] for group in raw_groups]
-    else:
-        groups = [[extraction_dir / work_id]]
-
-    collation_area = args.destdir / 'site' / 'images'
-    collation_area.mkdir(parents=True, exist_ok=True)
-
-    work_collation_dir = collation_area / work_id
-    if work_collation_dir.exists():
-        if len(list(work_collation_dir.iterdir())) > 0:
-            print('Collation directory already exists!')
-            return
-        else:
-            work_collation_dir.rmdir()
-
-    nonexistent = [path for group in (groups + [exclusions]) for path in group if not path.exists()]
-    if len(nonexistent) > 0:
-        print(f'Nonexistent paths: {nonexistent}')
-        return
-
-    collation_staging_area = args.destdir / 'site' / 'images-staging'
-    work_staging_dir = collation_staging_area / work_id
-    work_staging_dir.mkdir(parents=True)
-
-    collator = Collator(work_staging_dir, exclusions, args)
-    for group in groups:
-        collation_result = collator.collate_from_paths([item for item in group if item not in exclusions])
-        if collation_result is None:
-            print(f'Unable to deduce file structure for {work_id} subgroup {[str(path) for path in group]}')
-            break
-
-    if collation_result and collator.index > 0:
-        print(f'Collated {collator.index} pages for {work_id}')
-        work_staging_dir.rename(work_collation_dir)
-    else:
-        for f in work_staging_dir.iterdir():
-            f.unlink()
-        work_staging_dir.rmdir()
-
-        if collation_result and collator.index == 0:
-            print(f'No files found for {work_id}')
-
-    collation_staging_area.rmdir()
-
 
 def fmt_size(s):
     return f'{s[0]}x{s[1]}px'
@@ -1022,9 +1025,9 @@ argparser = argparse.ArgumentParser(
       subfolder.
     - `fetch` metadata and thumbnail images for extracted works
       from DLSite.
-    - `collate` and/or `manual-collate` extracted works,
+    - `collate` extracted works, producing a single sequence of
-      producing a single sequence of image files (or symlinks
+      image files (or symlinks into the extracted data, when
-      into the extracted data, when possible) for each work.
+      possible) for each work.
     - Manually adjust works' `metadata` when necessary.
     - `generate` a static website providing a catalog and viewer
       for all collated works.
@@ -1047,7 +1050,7 @@ argparser.add_argument(
 )
 subparsers = argparser.add_subparsers(title="subcommands", required=True)
 
-parser_extract = subparsers.add_parser('extract', aliases=['x', 'ex'], help='extract zipfiles')
+parser_extract = subparsers.add_parser('extract', aliases=['x'], help='extract zipfiles')
 parser_extract.add_argument(
     '-r', '--remove',
     action='store_true',
@@ -1062,111 +1065,96 @@ parser_extract.add_argument(
 )
 parser_extract.set_defaults(func=extract)
 
-parser_fetch = subparsers.add_parser('fetch', aliases=['f', 'fet'], help='fetch metadata and thumbnails')
+parser_fetch = subparsers.add_parser('fetch', aliases=['f'], help='fetch metadata and thumbnails')
 parser_fetch.set_defaults(func=fetch)
 
 parser_collate = subparsers.add_parser(
     'collate',
-    aliases=['c', 'co', 'col'],
+    aliases=['c'],
-    help='collate each work into a sequence of image files',
+    help='collate works into sequences of image files',
     formatter_class=argparse.RawDescriptionHelpFormatter,
     description=textwrap.dedent("""\
     For each extracted work that has not already been collated,
-    DLibrary will attempt to intuit its structure as follows:
+    DLibrary will attempt to intuit its structure and create
+    a single ordered list of image files in the site data
+    directory. Each image will either be a symlink to an image
+    file in the extraction folder, or a single page extracted
+    from a PDF file.
 
-    - Enter the work's directory. If the directory contains
+    DLibrary may fail to automatically collate a work if its
-      nothing except a single subdirectory (ignoring a few types
+    files and subdirectories are not named in a way that
-      of files that are definitely not relevant), traverse
+    indicates a clear linear ordering. In order to assist with
-      downwards repeatedly.
+    collation, you can provide a list of expressions specifying
-    - If the current directory contains nothing except a single
+    where to start traversing the directory structure, what
-      PDF (again, ignoring irrelevant files), attempt to extract
+    files to include in what order, and/or what files to ignore
-      a series of images from the PDF. This process expects that
+    entirely.
-      each page of the PDF consists of a single embedded image,
-      which will be extracted at full resolution. Support for
-      more complex PDFs is not yet implemented.
-    - If the current directory contains nothing except image
-      files, and the image files are named in a way that clearly
-      indicates a complete numerical order (each filename
-      consists of a shared prefix followed by a distinct
-      number), symlink files in the inferred order.
-    - Otherwise, skip processing this work for now.
 
-    DLibrary can be given "collation hints" which provide
+    An expression can be:
-    alternative starting points for this search process. A hint
-    is a path under $DLIBRARY_DIR/extract/[work id]/
-    indicating a different directory or PDF file to begin the
-    search process for that work, rather than starting at the
-    top level of the extracted data. There can be at most one
-    hint per work; for more complicated scenarios where a work
-    includes multiple folders that need to be collated together,
-    or where filenames do not clearly indicate an ordering, use
-    `manual-collate` instead.
-    """),
-)
-parser_collate.add_argument(
-    'hints',
-    metavar='PATH',
-    type=Path,
-    nargs='*',
-    help='paths within extraction folders as collation hints'
-)
-parser_collate.set_defaults(func=collate, force_convert_pdf=False)
-
-parser_manual_collate = subparsers.add_parser(
-    'manual-collate',
-    aliases=['mc', 'man', 'manual'],
-    help='collate a single work manually',
-    formatter_class=argparse.RawDescriptionHelpFormatter,
-    description=textwrap.dedent("""\
-    Provide an expression or sequence of expressions specifying groups
-    of paths to collate or skip. An expression can be:
 
     PATH
       A single path. If this is an image, it will be appended to
-      the sequence of collated images; if this is a PDF, images will be
+      the sequence of collated images for the work it belongs to;
-      extracted from it and concatenated to the sequence; if this is a
+      if this is a PDF, images will be extracted from it and
-      directory, the contents of the directory will be collated based on
+      concatenated to the sequence; if this is a directory, the
-      the normal heuristics and concatenated to the sequence.
+      contents of the directory will be automatically collated
+      using DLibrary's default heuristics, and concatenated
+      to the sequence.
 
     ( PATH [PATH ...] )
       A group of paths contained in parentheses. You may need to escape
       the parentheses to avoid them getting parsed by your shell.
       All the paths in this group will be considered together, and
-      collated based on the normal heuristics, regardless of what
+      automatically collated using the default heuristics, regardless
-      order the paths are provided in.
+      of what order the paths are provided in.
 
     ! PATH
     ! ( PATH [PATH ...] )
       A path or group of paths to exclude from collation. You may
       need to escape the !. If an excluded path appears within any
-      of the other specified paths, it will be ignored.
+      of the other specified paths, it will be skipped by the collation
+      heuristics.
 
     If the only expressions provided are negations, then auto-collation
-    will start from the top level of the extracted work while excluding
+    will start from the top level of the extracted work while skipping
-    the negated paths.
+    the excluded paths.
 
     All provided paths must be under $DLIBRARY_DIR/extract/[work id]/
-    for the work being manually collated. `manual-collate` can
+    for some not-yet-collated work. Paths belonging to multiple
-    only handle one work at a time.
+    different works can all be provided on the same command line, and
+    expressions will be clustered together by work id while otherwise
+    preserving the order they were provided in. A parenthesized group
+    expression must only contain paths belonging to a single work.
+
+    By default, DLibrary will attempt to collate every not-yet-collated
+    work (excluding "virtual" works), using the provided expressions
+    to assist in collation when available. The `-o` flag will direct
+    DLibrary to *only* collate works included in the provided expressions,
+    even if other uncollated works are present.
     """),
 )
-parser_manual_collate.add_argument(
+parser_collate.add_argument(
+    '-o', '--only-specified-works',
+    action='store_true',
+    help="only collate works that are explicitly specified",
+)
+parser_collate.add_argument(
     '--force-convert-pdf',
     action='store_true',
     help="convert a PDF page to a 300dpi image if there isn't a single image we can extract directly",
 )
-parser_manual_collate.add_argument(
+parser_collate.add_argument(
     'expression',
-    nargs='+',
+    nargs='*',
     help='expressions indicating paths to collate or skip',
 )
-parser_manual_collate.set_defaults(func=manual_collate)
+parser_collate.set_defaults(func=collate)
 
-parser_analyze = subparsers.add_parser('analyze', aliases=['a', 'an', 'anal'], help='analyze an extracted folder to assist in collation')
+
+parser_analyze = subparsers.add_parser('analyze', aliases=['a'], help='analyze an extracted folder to assist in collation')
 parser_analyze.add_argument('work_id')
 parser_analyze.set_defaults(func=analyze)
 
-parser_metadata = subparsers.add_parser('metadata', aliases=['m', 'me', 'meta'], help='view or modify metadata for a work')
+parser_metadata = subparsers.add_parser('metadata', aliases=['m'], help='view or modify metadata for a work')
 parser_metadata.add_argument('work_id')
 parser_metadata.add_argument(
     '--virtual',
@@ -1177,7 +1165,7 @@ parser_metadata.set_defaults(func=metadata)
 
 parser_generate = subparsers.add_parser(
     'generate',
-    aliases=['g', 'gen'],
+    aliases=['g'],
     help='generate HTML/CSS/JS for library site',
     formatter_class=argparse.RawDescriptionHelpFormatter,
     description=textwrap.dedent("""\