Skip to content

Commit

Permalink
cache: If original_file_name has changed delete old entry
Browse files Browse the repository at this point in the history 
Make cache exceptions non fatal

If the cache fails for whatever reason then carry on without it. We
generally don't want cache errors to be fatal but they are useful to
know about so we just print the exceptions which will get picked up by
the monitoring.

Fixes: #48
  • Loading branch information
@michaelwood
michaelwood committed Oct 2, 2023
1 parent 7abdfd4 commit 7b710d8
Show file tree
Hide file tree
Showing 2 changed files with 102 additions and 70 deletions.
121 changes: 71 additions & 50 deletions getter/cache.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,77 +6,98 @@
DATABASE_NAME = "cache_datagetter.db"
CACHE_DIR = "cache_dir"

class DatagetterCacheError(Exception):
pass


def setup_database():
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()
cur.execute(
"""CREATE TABLE IF NOT EXISTS cache
(original_file_name TEXT NOT NULL UNIQUE,
hash TEXT NOT NULL UNIQUE,
json_file TEXT NOT NULL UNIQUE);"""
)
con.commit()
con.close()
try:
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()
cur.execute(
"""CREATE TABLE IF NOT EXISTS cache
(original_file_name TEXT NOT NULL UNIQUE,
hash TEXT NOT NULL UNIQUE,
json_file TEXT NOT NULL UNIQUE);"""
)
con.commit()
con.close()
except Exception as e:
raise DatagetterCacheError(e)


def setup_cache_dir():
os.makedirs(CACHE_DIR, exist_ok=True)
try:
os.makedirs(CACHE_DIR, exist_ok=True)
except Exception as e:
raise DatagetterCacheError(e)


def hash_file(original_file_path):
file_hash = hashlib.sha1()
try:
file_hash = hashlib.sha1()

with open(original_file_path, "rb") as fp:
while True:
data = fp.read(6000)
if not data:
break
with open(original_file_path, "rb") as fp:
while True:
data = fp.read(6000)
if not data:
break

file_hash.update(data)
file_hash.update(data)
except Exception as e:
raise DatagetterCacheError(e)

return file_hash.hexdigest()


def get_file(file_hash_str):
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()
cur.execute("SELECT json_file FROM cache WHERE hash = ?", (file_hash_str,))
row = cur.fetchone()
con.close()
try:
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()
cur.execute("SELECT json_file FROM cache WHERE hash = ?", (file_hash_str,))
row = cur.fetchone()
con.close()

# We haven't already converted the file
if not row:
return False
# We haven't already converted the file
if not row:
return False

# We have converted the file before so copy from the CACHE_DIR
return os.path.join(CACHE_DIR, row[0])
# We have converted the file before so copy from the CACHE_DIR
return os.path.join(CACHE_DIR, row[0])
except Exception as e:
raise DatagetterCacheError(e)


def update_cache(json_file_name, file_hash_str, file_identifier, file_type):
"""
Updates the cache database and copies the data into the cache dir
json_file_name: Output desination for the file
"""
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()

cur.execute(
"""
INSERT INTO cache (original_file_name, hash, json_file)
VALUES (?, ?, ?)
ON CONFLICT(original_file_name) DO UPDATE SET hash=?
WHERE original_file_name=?
""",
(
file_identifier + "." + file_type,
file_hash_str,
file_identifier + ".json",
file_hash_str,
file_identifier + "." + file_type,
),
)

con.commit()

shutil.copy(json_file_name, CACHE_DIR)
# Todo clean up cache functionality for orphaned files or to reset the cache

try:
con = sqlite3.connect(DATABASE_NAME)
cur = con.cursor()

cur.execute(
"""
INSERT INTO cache (original_file_name, hash, json_file)
VALUES (?, ?, ?)
ON CONFLICT(original_file_name) DO UPDATE SET hash=?
ON CONFLICT(json_file) DO UPDATE SET hash=?, original_file_name=?
""",
(
file_identifier + "." + file_type,
file_hash_str,
file_identifier + ".json",
file_hash_str,
file_hash_str,
file_identifier + "." + file_type,
),
)

con.commit()

shutil.copy(json_file_name, CACHE_DIR)
except Exception as e:
raise DatagetterCacheError(e)
51 changes: 31 additions & 20 deletions getter/get.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -211,18 +211,22 @@ def fetch_and_convert(args, dataset, schema_path, package_schema):
print("Running convert on %s to %s" % (original_file_path,
json_file_name))

# Hash the file
file_hash_str = cache.hash_file(original_file_path)

# Check if we have already converted the file
cached_file_path = cache.get_file(file_hash_str)

# We have converted the file before so copy from the CACHE_DIR
if cached_file_path:
try:
shutil.copy(cached_file_path, json_file_name)
print("Cache hit")
except FileNotFoundError:
try:
# Hash the file
file_hash_str = cache.hash_file(original_file_path)
# Check if we have already converted the file
cached_file_path = cache.get_file(file_hash_str)

# We have converted the file before so copy from the CACHE_DIR
if cached_file_path:
try:
shutil.copy(cached_file_path, json_file_name)
print("Cache hit")
except (FileNotFoundError, PermissionError):
cached_file_path = False
except cache.DatagetterCacheError as e:
print(e)
print("Continuing without cache")
cached_file_path = False

if not cached_file_path:
Expand All @@ -233,12 +237,15 @@ def fetch_and_convert(args, dataset, schema_path, package_schema):
schema_path,
args.schema_branch
)

cache.update_cache(
json_file_name,
file_hash_str,
dataset['identifier'],
file_type)
try:
cache.update_cache(
json_file_name,
file_hash_str,
dataset['identifier'],
file_type)
except cache.DatagetterCacheError as e:
print(e)
print("Continuing without cache")

except KeyboardInterrupt:
raise
Expand Down Expand Up @@ -309,8 +316,12 @@ def file_cache_schema(schema_branch):

def get(args):

cache.setup_database()
cache.setup_cache_dir()
try:
cache.setup_database()
cache.setup_cache_dir()
except cache.DatagetterCacheError as e:
print(e)
print("Continuing without cache")

if not args.download:
mkdirs(args.data_dir, True)
Expand Down

0 comments on commit 7b710d8

Please sign in to comment.