Spaces:
Sleeping
Sleeping
Commit
·
e978f7c
1
Parent(s):
50a6a65
Add log2duckdb
Browse files- utils/log2duckdb.py +88 -0
utils/log2duckdb.py
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import duckdb
|
| 2 |
+
from dateutil.parser import parse
|
| 3 |
+
|
| 4 |
+
from config.log_definitions import log_definitions
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
class LogParser:
|
| 8 |
+
"""
|
| 9 |
+
A class that takes a log file path and a log definition (for example from log_definitions),
|
| 10 |
+
then parses the file and returns a DuckDB relation containing the extracted data.
|
| 11 |
+
"""
|
| 12 |
+
|
| 13 |
+
def __init__(self, file_path, log_type, db_path=":memory:"):
|
| 14 |
+
self.file_path = file_path
|
| 15 |
+
self.log_definition = log_definitions[log_type]
|
| 16 |
+
self.log_separator = self.log_definition["sep"]
|
| 17 |
+
self.conn = duckdb.connect(db_path)
|
| 18 |
+
self.table_name = f"log_{log_type}"
|
| 19 |
+
|
| 20 |
+
def parse_line(self, line):
|
| 21 |
+
"""Parse a line from the log file using the provided definition."""
|
| 22 |
+
# Start by replacing [ and ] with spaces
|
| 23 |
+
line = line.replace("[", " ").replace("]", " ")
|
| 24 |
+
|
| 25 |
+
# Get separator from log definition, default to whitespace if not specified
|
| 26 |
+
tokens = line.strip().split(self.log_separator)
|
| 27 |
+
|
| 28 |
+
# Ignore the line if it does not contain enough tokens
|
| 29 |
+
if len(tokens) < len(self.log_definition["fields"]):
|
| 30 |
+
return None
|
| 31 |
+
|
| 32 |
+
entry = {}
|
| 33 |
+
for field in self.log_definition["fields"]:
|
| 34 |
+
pos = field["pos"]
|
| 35 |
+
|
| 36 |
+
# Extract the value according to the indicated position
|
| 37 |
+
if isinstance(pos, slice):
|
| 38 |
+
value = " ".join(tokens[pos])
|
| 39 |
+
else:
|
| 40 |
+
try:
|
| 41 |
+
value = tokens[pos]
|
| 42 |
+
except IndexError:
|
| 43 |
+
value = None
|
| 44 |
+
|
| 45 |
+
# Type conversion
|
| 46 |
+
if "type" in field:
|
| 47 |
+
typ = field["type"]
|
| 48 |
+
if typ == "datetime":
|
| 49 |
+
# Try to parse the date with dateutil.parser
|
| 50 |
+
try:
|
| 51 |
+
value = parse(value)
|
| 52 |
+
except ValueError:
|
| 53 |
+
# If the date is not parsable, try several formats
|
| 54 |
+
value = None
|
| 55 |
+
elif typ == "direction":
|
| 56 |
+
value = "download" if value == "o" else "upload"
|
| 57 |
+
else:
|
| 58 |
+
try:
|
| 59 |
+
value = typ(value)
|
| 60 |
+
except Exception:
|
| 61 |
+
pass
|
| 62 |
+
|
| 63 |
+
entry[field["name"]] = value
|
| 64 |
+
|
| 65 |
+
return entry
|
| 66 |
+
|
| 67 |
+
def parse_file(self):
|
| 68 |
+
"""Iterate through the entire log file and return a DuckDB relation containing the parsed entries."""
|
| 69 |
+
data = []
|
| 70 |
+
with open(self.file_path, "r") as f:
|
| 71 |
+
for line in f:
|
| 72 |
+
parsed = self.parse_line(line)
|
| 73 |
+
if parsed:
|
| 74 |
+
data.append(parsed)
|
| 75 |
+
|
| 76 |
+
if not data:
|
| 77 |
+
return None
|
| 78 |
+
|
| 79 |
+
# Create a DuckDB relation directly from the list of dictionaries
|
| 80 |
+
relation = self.conn.execute("SELECT * FROM ?", [data]).fetch_arrow_table()
|
| 81 |
+
|
| 82 |
+
# Create a table from the relation
|
| 83 |
+
self.conn.execute(
|
| 84 |
+
f"CREATE OR REPLACE TABLE {self.table_name} AS SELECT * FROM relation"
|
| 85 |
+
)
|
| 86 |
+
|
| 87 |
+
# Return a DuckDB relation
|
| 88 |
+
return self.conn.table(self.table_name)
|