smeserver-mailstats/root/usr/bin/mailstats.py

649 lines
24 KiB
Python

#
# Mailstats.py
#
#
# This script provides daily SpamFilter statistics.
#
# Re-written in python from Mailstats.pl (Perl) to conform to SME11 / Postfix / qpsmtpd log formats
# and html output added
#
# Todo
# 1. Make "yesterday" parameterised
# 2. Get data from SMEServer DB records
# 3. Other stats
# 4. Percentages for main table
# 5. Percentages and sort for Sub tables
#
from datetime import datetime, timedelta
import sys
from chameleon import PageTemplateFile,PageTemplate
import pkg_resources
import re
import ipaddress
import subprocess
import os
from collections import defaultdict
Mailstats_version = '1.2'
script_dir = os.path.dirname(os.path.abspath(__file__))
data_file_path = script_dir+'/../../../'
now = datetime.now()
yesterday = now - timedelta(days=1)
formatted_yesterday = yesterday.strftime("%Y-%m-%d")
# Column numbering
Hour = 0
WebMail = 1
Local = 2
MailMan = 3
Relay = 4
DMARC = 5
Virus = 6
RBLDNS = 7
Geoip = 8
NonConf = 9
RejLoad = 10
Karma = 11
DelSpam = 12
QuedSpam = 13
Ham = 14
TOTALS = 15
PERCENT = 16
ColTotals = 24
def is_private_ip(ip):
try:
# Convert string to an IPv4Address object
ip_addr = ipaddress.ip_address(ip)
except ValueError:
return False
# Define private IP ranges
private_ranges = [
ipaddress.ip_network('10.0.0.0/8'),
ipaddress.ip_network('172.16.0.0/12'),
ipaddress.ip_network('192.168.0.0/16'),
]
# Check if the IP address is within any of these ranges
for private_range in private_ranges:
if ip_addr in private_range:
return True
return False
def truncate_microseconds(timestamp):
# Split timestamp into main part and microseconds
try:
main_part, microseconds = timestamp.split('.')
# Truncate the last three digits of the microseconds
truncated_microseconds = microseconds[:-3]
# Combine the main part and truncated microseconds
truncated_timestamp = f"{main_part}.{truncated_microseconds}"
except Exception as e:
print(f"{e} {timestamp}")
raise ValueError
# Remove the microseconds completely if they exist
return truncated_timestamp.split('.')[0]
# def filter_yesterdays_entries(log_entries):
# # Determine yesterday's date
# yesterday = (datetime.datetime.now() - datetime.timedelta(days=1)).date()
# # Filter entries for yesterday's date
# yesterday_entries = []
# for timestamp, data in log_entries:
# truncated_timestamp = truncate_microseconds(timestamp)
# entry_date = datetime.datetime.strptime(truncated_timestamp, '%Y-%m-%d %H:%M:%S').date()
# if entry_date == yesterday:
# parsed_data = parse_data(data)
# yesterday_entries.append((truncated_timestamp, parsed_data))
# return yesterday_entries
def read_in_yesterday_log_file(file_path):
# Read the file and split each line into a list - timestamp and the rest
# Get current date and calculate yesterday's date
log_entries = []
skip_record_count = 0;
with open(file_path, 'r') as file:
for Line in file:
#extract time stamp
try:
entry = split_timestamp_and_data(Line)
# compare with yesterday
timestamp_str = truncate_microseconds(entry[0])
except ValueError as e:
#print(f"ValueError {e} on timestamp create {timestamp_str}:{entry[0]} {entry[1]}")
skip_record_count += 1
continue
# Parse the timestamp string into a datetime object
# Ignoring extra microseconds
try:
timestamp = datetime.strptime(timestamp_str, "%Y-%m-%d %H:%M:%S")
except ValueError as e:
print(f"ValueError {e} on timestamp extract {timestamp_str}:{entry[1]}")
if timestamp.date() == yesterday.date():
log_entries.append((timestamp, entry[1]))
return [log_entries,skip_record_count]
def filter_summary_records(log_entries):
# Return just the summary records
filtered_log_entries = []
skipped_entry_count = 0
for line in log_entries:
#print(line)
#quit()
if '`' in line[1]:
filtered_log_entries.append(line)
else:
skipped_entry_count += 1
return [filtered_log_entries,skipped_entry_count]
def sort_log_entries(log_entries):
# Sort the records, based on the timestamp
sorted_entries = sorted(log_entries, key=lambda x: x[0])
# and return a dictionary
sorted_dict = {entry[0]: entry[1] for entry in sorted_entries}
return sorted_dict
# def read_and_filter_yesterday_log(file_path):
# # Read the file and split each line into a dictionary
# log_entries = []
# with open(file_path, 'r') as file:
# for line in file:
# if '`' in line:
# parts = line.split(' ')
# if parts:
# # Combine parts to form the complete timestamp
# timestamp = ' '.join(parts[:2])
# data = ' '.join(parts[2:]) # The rest of the line after date and time
# log_entries.append((timestamp, data))
# # Filter the entries to keep only those from yesterday
# filtered_entries = filter_yesterdays_entries(log_entries)
# # Sort the filtered log entries based on the truncated timestamp
# sorted_entries = sorted(filtered_entries, key=lambda x: datetime.datetime.strptime(x[0], '%Y-%m-%d %H:%M:%S'))
# # Create a dictionary
# sorted_dict = {entry[0]: entry[1] for entry in sorted_entries}
# return sorted_dict
def parse_data(data):
# Split data string into parts and map to named fields.
# Adjust the field names and parsing logic according to your data format.
# Split at the backtick - before it fields split at space, after, fields split at tab
parts = data.split('`')
#print(f"{parts[0]}:{parts[1]}")
fields1 = parts[0].strip().split() if len(parts) > 0 else []
fields2 = parts[1].split('\t') if len(parts) > 1 else []
# then merge them
fields = fields1 + fields2
# if fields[8] != 'queued':
# i = 0
# print(f"len:{len(fields)}")
# for part in fields:
# print(f"{i}: {part}")
# i = i +1
# quit()
# and mapping:
try:
return_dict = {
'id': fields[0].strip() if len(fields) > 0 else None,
'action': fields[1].strip() if len(fields) > 1 else None,
'logterse': fields[2].strip() if len(fields) > 2 else None,
'ip': fields[3].strip() if len(fields) > 3 else None,
'sendurl': fields[4].strip() if len(fields) > 4 else None, #1
'sendurl1': fields[5].strip() if len(fields) > 5 else None, #2
'from-email': fields[6].strip() if len(fields) > 6 else None, #3
'error-reason': fields[6].strip() if len(fields) > 6 else None, #3
'to-email': fields[7].strip() if len(fields) > 7 else None, #4
'error-plugin': fields[8].strip() if len(fields) > 8 else None, #5
'action1': fields[8].strip() if len(fields) > 8 else None, #5
'error-number' : fields[9].strip() if len(fields) > 9 else None, #6
'sender': fields[10].strip() if len(fields) > 10 else None, #7
'error-msg' :fields[10].strip() if len(fields) > 10 else None, #7
'spam-status': fields[11].strip() if len(fields) > 11 else None, #8
'error-result': fields[11].strip() if len(fields) > 11 else None,#8
# Add more fields as necessary
}
except:
#print(f"error:len:{len(fields)}")
return_dict = {}
return return_dict
def count_entries_by_hour(log_entries):
hourly_counts = defaultdict(int)
for entry in log_entries:
# Extract hour from the timestamp
timestamp = entry['timestamp']
hour = datetime.datetime.strptime(timestamp, '%Y-%m-%d %H:%M:%S').strftime('%Y-%m-%d %H')
hourly_counts[hour] += 1
return hourly_counts
def initialize_2d_array(num_hours, column_headers_len,reporting_date):
num_hours += 1 # Adjust for the zeroth hour
# Initialize the 2D list with zeroes
return [[0] * column_headers_len for _ in range(num_hours)]
def search_2d_list(target, data):
"""
Search for a target string in a 2D list of variable-length lists of strings.
:param target: str, the string to search for
:param data: list of lists of str, the 2D list to search
:return: int, the row number where the target string is found, or -1 if not found
"""
for row_idx, row in enumerate(data):
if target in row:
return row_idx
return -1 # Return -1 if not found
def check_html2text_installed():
try:
# Check if html2text is installed by running 'which html2text'
result = subprocess.run(
['which', 'html2text'],
check=True,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE
)
# If the command finds html2text, it will output the path
html2text_path = result.stdout.decode('utf-8').strip()
if not html2text_path:
raise FileNotFoundError
print(f"html2text is installed at: {html2text_path}")
return True
except subprocess.CalledProcessError:
print("html2text is not installed. Please install it using your package manager.", file=sys.stderr)
return False
def html_to_text(input_file, output_file):
if not check_html2text_installed():
sys.exit(1)
try:
# Run the html2text command with -b0 --pad-tables parameters
result = subprocess.run(
['html2text', '-b0', '--pad-tables', input_file],
check=True, # Raise a CalledProcessError on non-zero exit
stdout=subprocess.PIPE, # Capture stdout
stderr=subprocess.PIPE # Capture stderr
)
# Write the stdout from the command to the output file
with open(output_file, 'w', encoding='utf-8') as outfile:
outfile.write(result.stdout.decode('utf-8'))
print(f"Converted {input_file} to {output_file}")
except subprocess.CalledProcessError as e:
print(f"Error occurred: {e.stderr.decode('utf-8')}", file=sys.stderr)
sys.exit(e.returncode)
def get_html2text_version():
try:
result = subprocess.run(['html2text', '--version'], stdout=subprocess.PIPE, stderr=subprocess.PIPE, universal_newlines=True)
# Ensure the result is treated as a string in Python 3.6+
return result.stdout.strip()
except subprocess.CalledProcessError as e:
print(f"Error occurred while checking html2text version: {e}", file=sys.stderr)
return None
def print_progress_bar(iteration, total, prefix='', suffix='', decimals=1, length=50, fill='', print_end="\r"):
"""
Call in a loop to create a terminal progress bar
@params:
iteration - Required : current iteration (Int)
total - Required : total iterations (Int)
prefix - Optional : prefix string (Str)
suffix - Optional : suffix string (Str)
decimals - Optional : positive number of decimals in percent complete (Int)
length - Optional : character length of bar (Int)
fill - Optional : bar fill character (Str)
print_end - Optional : end character (e.g. "\r", "\r\n") (Str)
"""
if total == 0:
raise ValueError("Progress total is zero")
percent = ("{0:." + str(decimals) + "f}").format(100 * (iteration / float(total)))
filled_length = int(length * iteration // total)
bar = fill * filled_length + '-' * (length - filled_length)
print(f'\r{prefix} |{bar}| {percent}% {suffix}', end=print_end)
# Print New Line on Complete
if iteration == total:
print()
def insert_string_after(original:str, to_insert:str, after:str) -> str:
"""
Insert to_insert into original after the first occurrence of after.
:param original: The original string.
:param to_insert: The string to be inserted.
:param after: The set of characters after which the string will be inserted.
:return: The new string with to_insert inserted after after.
"""
position = original.find(after)
#print(position)
if position == -1:
# 'after' string is not found in 'original'
return original
#print(f"{len(after)}")
# Position of the insertion point
insert_pos = position + len(after)
return original[:insert_pos] + to_insert + original[insert_pos:]
def split_timestamp_and_data(log_entry: str) -> list:
"""
Split a log entry into timestamp and the rest of the data.
:param log_entry: The log entry as a string.
:return: A list with two entries: [timestamp, rest_of_data].
"""
# The timestamp is always the first part, up to the first space after the milliseconds
parts = log_entry.split(' ', 2)
if len(parts) < 3:
raise ValueError(f"The log entry format is incorrect {parts}")
timestamp = ' '.join(parts[:2])
rest_of_data = parts[2]
#print(f"{timestamp} {rest_of_data}")
return [timestamp, rest_of_data]
def render_sub_table(table_title,table_headers,found_values):
# NeedNOTE: also need to compute the percentages here.
template_path = data_file_path+'mailstats-sub-table.html.pt'
# Load the template
with open(template_path, 'r') as template_file:
template_content = template_file.read()
# Create a Chameleon template instance
try:
template = PageTemplate(template_content)
# Render the template with the 2D array data and column headers
try:
rendered_html = template(array_2d=found_values, column_headers=table_headers, title=table_title)
except Exception as e:
print(f"{table_title}: A chameleon controller render error occurred: {e}")
raise ValueError
except Exception as e:
print(f"{table_title}: A chameleon controller template error occurred: {e}")
raise ValueError
return rendered_html
if __name__ == "__main__":
try:
chameleon_version = pkg_resources.get_distribution("Chameleon").version
except pkg_resources.DistributionNotFound:
chameleon_version = "Version information not available"
python_version = sys.version
python_version = python_version[:8]
current_datetime = datetime.now()
formatted_datetime = current_datetime.strftime("%Y-%m-%d %H:%M")
#From SMEServer DB
DomainName = 'bjsystems.co.uk' # $cdb->get('DomainName')->value;
RHSenabled = True #( $cdb->get('qpsmtpd')->prop('RHSBL') eq 'enabled' );
DNSenabled = True #( $cdb->get('qpsmtpd')->prop('DNSBL') eq 'enabled' );
SARejectLevel = 12 #$cdb->get('spamassassin')->prop('RejectLevel');
SATagLevel = 4 #$cdb->get('spamassassin')->prop('TagLevel');
FetchmailIP = '127.0.0.200'; #Apparent Ip address of fetchmail deliveries
WebmailIP = '127.0.0.1'; #Apparent Ip of Webmail sender
localhost = 'localhost'; #Apparent sender for webmail
FETCHMAIL = 'FETCHMAIL'; #Sender from fetchmail when Ip address not 127.0.0.200 - when qpsmtpd denies the email
MAILMAN = "bounces"; #sender when mailman sending when orig is localhost
DMARCDomain="dmarc"; #Pattern to recognised DMARC sent emails (this not very reliable, as the email address could be anything)
DMARCOkPattern="dmarc: pass"; #Pattern to use to detect DMARC approval
hello_string = "Mailstats:"+Mailstats_version+' for '+DomainName+" at "+formatted_datetime+" for "+formatted_yesterday
print(hello_string)
version_string = "Chameleon:"+chameleon_version+" Python:"+python_version
print(version_string)
num_hours = 25 # Represents hours from 0 to 23 - adds extra one for column totals and another for percentages
data_file = data_file_path+'current.log'
log_entries,skip_count = read_in_yesterday_log_file(data_file)
if len(log_entries) == 0:
print(f"No records found in {data_file}")
quit()
else:
print(f"Found {len(log_entries)} entries in log for for {formatted_yesterday} skipped {skip_count}")
summary_log_entries,skip_count = filter_summary_records(log_entries)
print(f"Found {len(summary_log_entries)} summary entries and skipped {skip_count} entries")
sorted_log_dict = sort_log_entries(summary_log_entries)
print(f"Sorted {len(sorted_log_dict)} entries")
columnHeaders = ['Count','WebMail','Local','MailMan','Relay','DMARC','Virus','RBL/DNS','Geoip.','Non.Conf.','Karma','Rej.Load','Del.Spam','Qued.Spam?',' Ham','TOTALS','PERCENT']
# dict for each colum identifying plugin that increments count
columnPlugin = [''] * 17
columnPlugin[Hour] = []
columnPlugin[WebMail] = []
columnPlugin[Local] = []
columnPlugin[MailMan] = []
columnPlugin[DMARC] = ['dmarc']
columnPlugin[Virus] = ['pattern_filter', 'virus::pattern_filter','virus::clamav']
columnPlugin[RBLDNS] = ['rhsbl', 'dnsbl','uribl']
columnPlugin[Geoip] = ['check_badcountries']
columnPlugin[NonConf] = ['check_earlytalker','check_relay','check_norelay', 'require_resolvable_fromhost'
,'check_basicheaders','check_badmailfrom','check_badrcptto_patterns'
,'check_badrcptto','check_spamhelo','check_goodrcptto extn','rcpt_ok'
,'check_goodrcptto','check_smtp_forward','count_unrecognized_commands','tls','auth::auth_cvm_unix_local'
,'auth::auth_imap', 'earlytalker','resolvable_fromhost','relay','headers','mailfrom','badrcptto','helo'
,'check_smtp_forward','sender_permitted_from']
columnPlugin[RejLoad] = ['loadcheck']
columnPlugin[DelSpam] = []
columnPlugin[QuedSpam] = []
columnPlugin[Ham] = []
columnPlugin[TOTALS] = []
columnPlugin[PERCENT] = []
columnPlugin[Karma] = ['karma']
columnHeaders_len = len(columnHeaders)
columnCounts_2d = initialize_2d_array(num_hours, columnHeaders_len,formatted_yesterday)
virus_pattern = re.compile(r"Virus found: (.*)")
found_viruses = defaultdict(int)
found_qpcodes = defaultdict(int)
qpcodes_pattern = re.compile(r".*(\(.*\)).*'")
i = 0;
sorted_len= len(sorted_log_dict)
# Initial call to print the progress bar
print_progress_bar(0, sorted_len, prefix='Progress:', suffix='Complete', length=50)
for timestamp, data in sorted_log_dict.items():
i += 1
print_progress_bar(i, sorted_len, prefix='Scanning for main table:', suffix='Complete', length=50)
#print(f"{i*100/len}%")
# Count of in which hour it falls
#hour = datetime.datetime.strptime(timestamp, '%Y-%m-%d %H:%M:%S').strftime('%Y-%m-%d %H')
# Parse the timestamp string into a datetime object
dt = timestamp
hour = dt.hour
# parse the data
#print(data)
parsed_data = parse_data(data)
#print(f"parsed_data['action']:{parsed_data['action']}\n")
# Increment Count in which headings it falls
#Hourly count and column total
columnCounts_2d[hour][Hour] += 1
columnCounts_2d[ColTotals][Hour] += 1
#Row Totals
columnCounts_2d[hour][TOTALS] += 1
#Total totals
columnCounts_2d[ColTotals][TOTALS] += 1
#Queued email
if parsed_data['action'] == '(queue)':
columnCounts_2d[hour][Ham] += 1
columnCounts_2d[ColTotals][Ham] += 1
#spamassasin
if parsed_data['spam-status'].lower().startswith('yes'):
#Extract other parameters from this string
# example: Yes, score=10.3 required=4.0 autolearn=disable
spam_pattern = r'score=([\d.]+)\s+required=([\d.]+)'
match = re.search(spam_pattern, parsed_data['spam-status'])
if match:
score = float(match.group(1))
required = float(match.group(2))
#print(f"{parsed_data['spam-status']} / {score} {required}")
if score >= SARejectLevel:
columnCounts_2d[hour][DelSpam] += 1
columnCounts_2d[ColTotals][DelSpam] += 1
elif score >= required:
columnCounts_2d[hour][QuedSpam] += 1
columnCounts_2d[ColTotals][QuedSpam] += 1
#Local send
elif DomainName in parsed_data['sendurl']:
columnCounts_2d[hour][Local] += 1
columnCounts_2d[ColTotals][Local] += 1
#Relay or webmail
elif not is_private_ip(parsed_data['ip']) and is_private_ip(parsed_data['sendurl1']) and parsed_data['action1'] == 'queued':
#Relay
if parsed_data['action1'] == 'queued':
columnCounts_2d[hour][Relay] += 1
columnCounts_2d[ColTotals][Relay] += 1
elif WebmailIP in parsed_data['sendurl1'] and not is_private_ip(parsed_data['ip']):
#webmail
columnCounts_2d[hour][WebMail] += 1
columnCounts_2d[ColTotals][WebMail] += 1
elif localhost in parsed_data['sendurl']:
# but not if it comes from fetchmail
if not FETCHMAIL in parsed_data['sendurl1']:
# might still be from mailman here
if MAILMAN in parsed_data['sendurl1']:
#$mailmansendcount++;
#$localsendtotal++;
columnCounts_2d[hour][MailMan] += 1
columnCounts_2d[ColTotals][MailMan] += 1
#$counts{$abshour}{$CATMAILMAN}++;
#$localflag = 1;
else:
#Or sent to the DMARC server
#check for email address in $DMARC_Report_emails string
#my $logemail = $log_items[4];
if DMARCDomain in parsed_data['from-email']: #(index($DMARC_Report_emails,$logemail)>=0) or
#$localsendtotal++;
#$DMARCSendCount++;
localflag = 1;
else:
# ignore incoming localhost spoofs
if not 'msg denied before queued' in parsed_data['error-msg']:
#Webmail
#$localflag = 1;
#$WebMailsendtotal++;
columnCounts_2d[hour][WebMail] += 1
columnCounts_2d[ColTotals][WebMail] += 1
#$WebMailflag = 1;
else:
#$localflag = 1;
#$WebMailsendtotal++;
#$WebMailflag = 1;
columnCounts_2d[hour][WebMail] += 1
columnCounts_2d[ColTotals][WebMail] += 1
#Now increment the column which the plugin name indicates
if parsed_data['action'] == '(deny)' and parsed_data['error-plugin']:
#print(f"Found plugin {parsed_data['error-plugin']}")
if parsed_data['error-plugin']:
row = search_2d_list(parsed_data['error-plugin'],columnPlugin)
if not row == -1:
#print(f"Found row: {row}")
columnCounts_2d[hour][row] += 1
columnCounts_2d[ColTotals][row] += 1
# a few ad hoc extra extractons of data
if row == Virus:
match = virus_pattern.match(parsed_data['action1'])
if match:
found_viruses[match.group(1)] += 1
else:
found_viruses[parsed_data['action1']] += 1
elif parsed_data['error-plugin'] == 'naughty':
match = qpcodes_pattern.match(parsed_data['action1'])
if match:
rejReason = match.group(1)
found_qpcodes[parsed_data['error-plugin']+"-"+rejReason] += 1
else:
found_qpcodes['Unknown'] += 1
else:
found_qpcodes[parsed_data['action1']] += 1
print()
# Now scan for the other lines in the log of interest
found_countries = defaultdict(int)
geoip_pattern = re.compile(r".*check_badcountries: GeoIP Country: (.*)")
dmarc_pattern = re.compile(r".*dmarc: pass")
total_countries = 0
DMARCOkCount = 0
# Pick up all log_entries = read_yesterday_log_file(data_file)
sorted_log_dict = sort_log_entries(log_entries)
i = 0
sorted_len = len(sorted_log_dict)
print_progress_bar(0, sorted_len, prefix='Progress:', suffix='Complete', length=50)
for timestamp, data in sorted_log_dict.items():
i += 1
print_progress_bar(i, sorted_len, prefix='Scanning for sub tables:', suffix='Complete', length=50)
#Pull out Geoip countries for analysis table
match = geoip_pattern.match(data)
if match:
country = match.group(1)
found_countries[country] += 1
total_countries += 1
continue
#Pull out DMARC approvals
match = dmarc_pattern.match(data)
if match:
DMARCOkCount += 1
continue
#Now apply the results to the chameleon template - main table
# Path to the template file
template_path = data_file_path+'mailstats.html.pt'
# Load the template
with open(template_path, 'r') as template_file:
template_content = template_file.read()
# Create a Chameleon template instance
template = PageTemplate(template_content)
# Render the template with the 2D array data and column headers
rendered_html = template(array_2d=columnCounts_2d, column_headers=columnHeaders, reporting_date=formatted_yesterday, title=hello_string, version=version_string)
total_html = rendered_html
#Now apply the results to the chameleon template - subservient tables
#qpsmtd codes
qpsmtpd_headers = ["Code",'Count','Percent','Reason']
qpsmtpd_title = 'Qpsmtpd codes league table:'
rendered_html = render_sub_table(qpsmtpd_title,qpsmtpd_headers,found_qpcodes)
# Add it to the total
total_html = insert_string_after(total_html,rendered_html, "<!---Add in sub tables here -->")
#Geoip Country codes
geoip_headers = ['Country','Count','Percent','Rejected?']
geoip_title = 'Geoip results:'
rendered_html = render_sub_table(geoip_title,geoip_headers,found_countries)
# Add it to the total
total_html = insert_string_after(total_html,rendered_html, "<!---Add in sub tables here -->")
# Write the rendered HTML to a file
output_path = data_file_path+'mailstats_for_'+formatted_yesterday
output_path = output_path.replace(' ','_')
with open(output_path+'.html', 'w') as output_file:
output_file.write(total_html)
#and create a text version if the local version is suffiicent
if get_html2text_version() == '2019.9.26':
html_to_text(output_path+'.html',output_path+'.txt')
print(f"Rendered HTML saved to {output_path}.html/txt")