use locator db to speed up file search

[photosort.git] / src / locator.py

diff --git a/src/locator.py b/src/locator.py
new file mode 100644 (file)
index 0000000..55142c5
--- /dev/null
+++ b/src/locator.py
@@ -0,0 +1,74 @@
+import logging
+import os
+import sqlite3
+import time
+
+import misc
+
+class Locator:
+    def __init__(self, base_dir, exclude_dir, db_file):
+        self._base_dir = base_dir
+        self._exclude_dir = exclude_dir
+        if not os.path.isfile(db_file):
+            self._conn = sqlite3.connect(db_file)
+            self._create_db()
+        else:
+            self._conn = sqlite3.connect(db_file)
+        logging.info("Opened locator for %s.", self._base_dir)
+
+    def locate_file(self, name, size, meta_time):
+        c = self._conn.cursor()
+        c.execute("SELECT file_dir FROM cache WHERE name=? AND size=? AND meta_time=?",
+                  (name, size, meta_time))
+        cached = c.fetchone()
+        if cached:
+            if os.path.isfile(cached[0]) and \
+               os.path.getsize(cached[0]) == size and \
+               misc.extract_meta_time(cached[0]) == meta_time:
+                self._update_cache_access_time(name, size, meta_time)
+                return cached[0]
+            else:
+                self._remove_from_cache(name, size, meta_time)
+        else:
+            file_dir = misc.find_file(name, size, meta_time)
+            if file_dir:
+                self._clean_cache()
+                self._add_to_cache(name, size, meta_time, file_dir)
+            return file_dir
+
+    def close(self):
+        logging.info("Closing locator for %s...", self._base_dir)
+        self._conn.close()
+
+    def _update_cache_access_time(self, name, size, meta_time):
+        c = self._conn.cursor()
+        c.execute("UPDATE cache SET access_time=? WHERE name=? AND size=? AND meta_time=?",
+                  (int(time.time()), name, size, meta_time))
+        self._conn.commit()
+
+    def _remove_from_cache(self, name, size, meta_time):
+        c = self._conn.cursor()
+        c.execute("DELETE FROM cache WHERE name=? AND size=? AND meta_time=?",
+                  (name, size, meta_time))
+        self._conn.commit()
+
+    def _add_to_cache(self, name, size, meta_time, file_dir):
+        c = self._conn.cursor()
+        c.execute('''INSERT INTO cache (name, size, meta_time, file_dir, access_time)
+                     VALUES (?, ?, ?, ?, ?)''',
+                  (name, size, meta_time, file_dir, int(time.time())))
+        self._conn.commit()
+
+    def _clean_cache(self):
+        c = self._conn.cursor()
+        c.execute("DELETE FROM cache WHERE access_time<?", (int(time.time()) - 604800))
+        self._conn.commit()
+
+    def _create_db(self):
+        logging.info("Creating locator database for %s...", self._base_dir)
+        c = self._conn.cursor()
+        c.execute('''CREATE TABLE cache
+                       (name TEXT, size INTEGER, meta_time INTEGER, file_dir TEXT,
+                        access_time INTEGER)
+                     PRIMARY KEY (name, size, meta_time)''')
+        self._conn.commit()