mkm3u: Maintain a cache of durations because they take ages to look up.

diff --git a/.gitignore b/.gitignore
index 909c9bc..0940b00 100644 (file)
--- a/.gitignore
+++ b/.gitignore
@@ -1,2 +1,6 @@
 *.m3u8
+*.m3u8.new
 !/ref/*.m3u8
+
+/mkm3u.cache
+/mkm3u.cache-stamp

diff --git a/Makefile b/Makefile
index 4cf4bb6..33136ba 100644 (file)
--- a/Makefile
+++ b/Makefile
@@ -2,6 +2,7 @@
 
 all:
 clean::; rm -f $(CLEANFILES)
+realclean::; rm -f $(REALCLEANFILES)
 force:
 .PHONY: all clean
 .SECONDEXPANSION: # not sorry
@@ -17,6 +18,10 @@ v-tag.0                       = @$(call v-print.0,$1,$@);
 
 TARGETS                         =
 CLEANFILES             += $(TARGETS)
+REALCLEANFILES         += $(CLEANFILES)
+
+MKM3UFLAGS              = -dmkm3u.cache
+REALCLEANFILES         += mkm3u.cache
 
 define %declare-playlist
 PLAYLISTS              += $1
@@ -175,8 +180,17 @@ $(call declare-playlist, drwho-silurians, D/Doctor Who/S07E02 BBB. Doctor Who an
 M3US                    = $(addsuffix .m3u8,$(PLAYLISTS))
 TARGETS                        += $(M3US)
 
-$(M3US): %.m3u8: $$($$*_EPLS) mkm3u
-       $(call v-tag,MKM3U)./mkm3u $($*_MKM3UFLAGS) "$<" >"$@.new" && mv "$@.new" "$@"
+CLEANFILES             += mkm3u.cache-stamp
+mkm3u.cache-stamp:
+       if [ ! -f mkm3u.cache ]; then \
+         ./mkm3u -i -dmkm3u.cache.new && mv mkm3u.cache.new mkm3u.cache; \
+       fi
+       touch $@
+
+CLEANFILES             += *.m3u8.new
+$(M3US): %.m3u8: $$($$*_EPLS) mkm3u mkm3u.cache-stamp
+       $(call v-tag,MKM3U)./mkm3u $(MKM3UFLAGS) $($*_MKM3UFLAGS) \
+               "$<" >"$@.new" && mv "$@.new" "$@"
 
 CHECKS                  = $(foreach p,$(PLAYLISTS), check/$p)
 check: $(CHECKS)

diff --git a/mkm3u b/mkm3u
index 5464ac7..9053c5f 100755 (executable)
--- a/mkm3u
+++ b/mkm3u
@@ -2,9 +2,11 @@
 ### -*- mode: python; coding: utf-8 -*-
 
 from contextlib import contextmanager
+import errno as E
 import optparse as OP
 import os as OS
 import re as RX
+import sqlite3 as SQL
 import subprocess as SP
 import sys as SYS
 
@@ -97,6 +99,34 @@ class FileLocation (BaseLocation):
 
 LOC = DummyLocation()
 
+ROOT = "/mnt/dvd/archive/"
+DB = None
+
+def init_db(fn):
+  global DB
+  DB = SQL.connect(fn)
+  DB.cursor().execute("PRAGMA journal_mode = WAL")
+
+def setup_db(fn):
+  try: OS.unlink(fn)
+  except OSError as e:
+    if e.errno == E.ENOENT: pass
+    else: raise
+  init_db(fn)
+  DB.cursor().execute("""
+          CREATE TABLE duration
+                  (path TEXT NOT NULL,
+                   title INTEGER NOT NULL,
+                   start_chapter INTEGER NOT NULL,
+                   end_chapter INTEGER NOT NULL,
+                   inode INTEGER NOT NULL,
+                   device INTEGER NOT NULL,
+                   size INTEGER NOT NULL,
+                   mtime REAL NOT NULL,
+                   duration REAL NOT NULL,
+                   PRIMARY KEY (path, title, start_chapter, end_chapter));
+  """)
+
 class Source (object):
   PREFIX = ""
   TITLEP = CHAPTERP = False
@@ -128,7 +158,48 @@ class Source (object):
     else:
       suffix = "#%d:%d-%d:%d" % (title, start_chapter, title, end_chapter - 1)
 
-    duration = me._duration(title, start_chapter, end_chapter)
+    duration = None
+    if DB is None:
+      duration = me._duration(title, start_chapter, end_chapter)
+    else:
+      st = OS.stat(OS.path.join(ROOT, me.fn))
+      duration = None
+      c = DB.cursor()
+      c.execute("""
+              SELECT device, inode, size, mtime,  duration FROM duration
+              WHERE path = ? AND title = ? AND
+                    start_chapter = ? AND end_chapter = ?
+      """, [me.fn, title, start_chapter is None and -1 or start_chapter,
+            end_chapter is None and -1 or end_chapter])
+      row = c.fetchone()
+      foundp = False
+      if row is None:
+        duration = me._duration(title, start_chapter, end_chapter)
+        c.execute("""
+                INSERT OR REPLACE INTO duration
+                        (path, title, start_chapter, end_chapter,
+                         device, inode, size, mtime,  duration)
+                VALUES (?, ?, ?, ?,  ?, ?, ?, ?,  ?)
+        """, [me.fn, title, start_chapter is None and -1 or start_chapter,
+              end_chapter is None and -1 or end_chapter,
+              st.st_dev, st.st_ino, st.st_size, st.st_mtime,
+              duration])
+      else:
+        dev, ino, sz, mt,  d = row
+        if (dev, ino, sz, mt) == \
+           (st.st_dev, st.st_ino, st.st_size, st.st_mtime):
+          duration = d
+        else:
+          duration = me._duration(title, start_chapter, end_chapter)
+          c.execute("""
+                  UPDATE duration
+                  SET device = ?, inode = ?, size = ?, mtime = ?, duration = ?
+                  WHERE path = ? AND title = ? AND
+                        start_chapter = ? AND end_chapter = ?
+        """, [st.st_dev, st.st_dev, st.st_size, st.st_mtime,  duration,
+              me.fn, title, start_chapter is None and -1 or start_chapter,
+              end_chapter is None and -1 or end_chapter])
+      DB.commit()
 
     if end_chapter is not None:
       keys = [(title, ch) for ch in range(start_chapter, end_chapter)]
@@ -737,29 +808,41 @@ class EpisodeListParser (object):
                             (d.fn, d.neps, d.nuses))
     return me._pl
 
-ROOT = "/mnt/dvd/archive/"
-
 op = OP.OptionParser \
-  (usage = "%prog [-c] [-s SERIES] EPLS",
+  (usage = "%prog [-c] [-d CACHE] [-s SERIES] EPLS\n"
+           "%prog -i -d CACHE",
    description = "Generate M3U playlists from an episode list.")
 op.add_option("-c", "--chapters",
               dest = "chaptersp", action = "store_true", default = False,
               help = "Output individual chapter names")
+op.add_option("-i", "--init-db",
+              dest = "initdbp", action = "store_true", default = False,
+              help = "Initialize the database")
+op.add_option("-d", "--database",
+              dest = "database", type = "str", default = None,
+              help = "Set filename for cache database")
 op.add_option("-s", "--series",
               dest = "series", type = "str", default = None,
               help = "Output only the listed SERIES (comma-separated)")
-opts, argv = op.parse_args()
-if len(argv) != 1: op.print_usage(file = SYS.stderr); SYS.exit(2)
-if opts.series is None:
-  series_wanted = None
-else:
-  series_wanted = set()
-  for name in opts.series.split(","): series_wanted.add(name)
 try:
-  ep = EpisodeListParser(series_wanted, opts.chaptersp)
-  ep.parse_file(argv[0])
-  pl = ep.done()
-  pl.write(SYS.stdout)
+  opts, argv = op.parse_args()
+  if opts.initdbp:
+    if opts.chaptersp or opts.series is not None or \
+       opts.database is None or len(argv):
+      op.print_usage(file = SYS.stderr); SYS.exit(2)
+    setup_db(opts.database)
+  else:
+    if len(argv) != 1: op.print_usage(file = SYS.stderr); SYS.exit(2)
+    if opts.database is not None: init_db(opts.database)
+    if opts.series is None:
+      series_wanted = None
+    else:
+      series_wanted = set()
+      for name in opts.series.split(","): series_wanted.add(name)
+    ep = EpisodeListParser(series_wanted, opts.chaptersp)
+    ep.parse_file(argv[0])
+    pl = ep.done()
+    pl.write(SYS.stdout)
 except (ExpectedError, IOError, OSError) as e:
   LOC.report(e)
   SYS.exit(2)