# Copyright (c) 2010 Flowerfire, Inc.  All Rights Reserved.
urlscan = {

  # The name of the log format
  log.format.format_label = "URLScan Log Format"
  log.miscellaneous.log_data_type = "generic"  
  log.miscellaneous.log_format_type = "web_server"

  # The log is in this format if any of the first ten lines match this regular expression
  log.format.autodetect_regular_expression = "\\[[0-9]+-[0-9]+-[0-9]+ - [0-9]+:[0-9]+:[0-9]+\\] ----------------"

  # The format of dates and times in this log
  log.format.date_format = "mm/dd/yyyy"
  log.format.time_format = "hh:mm:ss"

  # All log field parsing will be done using the parsing filters
  log.format.parse_only_with_filters = "true"

  # Log fields
  log.fields = {

    date = {
      label = "$lang_stats.field_labels.date"
      type = "date"
      index = 0
      subindex = 0
      hierarchy_dividers = ""
      left_to_right = false
      leading_divider = "false"
    } # date

    time = {
      label = "$lang_stats.field_labels.time"
      type = "time"
      index = 0
      subindex = 0
      hierarchy_dividers = ""
      left_to_right = false
      leading_divider = "false"
    } # time

    source_ip = {
      label = "$lang_stats.field_labels.source_ip"
      type = "host"
      index = 0
      subindex = 0
    } # source_ip

    message = {
      label = "$lang_stats.field_labels.message"
      type = "flat"
      index = 0
      subindex = 0
    } # message

    site_instance = {
      label = "$lang_stats.field_labels.site_instance"
      type = "flat"
      index = 0
      subindex = 0
    } # site_instance

    raw_url = {
      label = "$lang_stats.field_labels.raw_url"
      type = "page"
      index = 0
      subindex = 0
      hierarchy_dividers = "/?"
      left_to_right = true
      leading_divider = "true"
    } # raw_url

  } # log.fields


  #          
  # Log Parsing Filters
  log.parsing_filters = {


    # Parse out date, time, source IP, message
    1 = {
      label = "1"
      comment = ""
      value = "collect_fields_using_regexp('^()\\\\[([0-9]+-[0-9]+-[0-9]+) - ([0-9]+:[0-9]+:[0-9]+)\\\\] Client at ([0-9.]*): (.*)$', '*KEY*,date,time,source_ip,message')"
    } # 1


    # Parse out message, Site Instance, Raw URL
    2 = {
      label = "2"
      comment = ""
      value = "collect_fields_using_regexp('()Client at [0-9.]*: (.*)  Site Instance=\\'([^\\']*)\\', Raw URL=\\'([^\\']*)\\'', '*KEY*,message,site_instance,raw_url')"
    } # 2


    # Accept this log entry
    3 = {
      label = "3"
      comment = ""
      value = "accept_collected_entry_using_regexp('^()', false)"
    } # 3


  } # log.parsing_filters


  # Database fields
  database.fields = {

    date_time = {
      label = "$lang_stats.field_labels.date_time"
      log_field = "date_time"
      type = "string"
      suppress_top = 0
      suppress_bottom = 3
      display_format_type = "date_time"
    } # date_time

    day_of_week = {
      label = "$lang_stats.field_labels.day_of_week"
      log_field = "day_of_week"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
      display_format_type = "day_of_week"
    } # day_of_week

    hour_of_day = {
      label = "$lang_stats.field_labels.hour_of_day"
      log_field = "hour_of_day"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
      display_format_type = "hour_of_day"
    } # hour_of_day

    message = {
      label = "$lang_stats.field_labels.message"
      log_field = "message"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
    } # message

    source_ip = {
      label = "$lang_stats.field_labels.source_ip"
      log_field = "source_ip"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
    } # source_ip

    raw_url = {
      label = "$lang_stats.field_labels.raw_url"
      log_field = "raw_url"
      type = "string"
      suppress_top = 0
      suppress_bottom = 9
    } # raw_url

    worm = {
      label = "$lang_stats.field_labels.worm"
      log_field = "worm"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
    } # worm

    domain_description = {
      label = "$lang_stats.field_labels.domain_description"
      log_field = "domain_description"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
    } # domain_description

    location = {
      label = "$lang_stats.field_labels.location"
      log_field = "location"
      type = "string"
      suppress_top = 0
      suppress_bottom = 3
    } # location

    site_instance = {
      label = "$lang_stats.field_labels.site_instance"
      log_field = "site_instance"
      type = "string"
      suppress_top = 0
      suppress_bottom = 2
    } # site_instance

  } # database.fields


  # Log Filters
  log.filters = {


    mark_entry = {
      label = '$lang_admin.log_filters.mark_entry_label'
      comment = '$lang_admin.log_filters.mark_entry_comment'
      value = 'hits = 1;'
    } # mark_entry

  } # log.filters

  log.field_options = {

    sessions_page_field = "raw_url"
    sessions_visitor_id_field = "source_ip"
    sessions_event_field = "page_views"

  } # log.field_options


  database.numerical_fields = {

    hits = {
      label = "$lang_stats.field_labels.hits"
      default = true
      requires_log_field = false
      type = "int"
      display_format_type = "integer"
      entries_field = true
    } # hits

    visitors = {
      label = "$lang_stats.field_labels.visitors"
      default = false
      requires_log_field = true
      log_field = "source_ip"
      type = "unique"
      display_format_type = "integer"
    } # visitors

  } # database.numerical_fields


  create_profile_wizard_options = {
    date_time_tracking = true

    # How the reports should be grouped in the report menu
    report_groups = {
      date_time_group = ""
      message = true
      source_ip = true
      raw_url = true
      worm = true
      domain_description = true
      location = true
      site_instance = true
    } # report_groups

  } # create_profile_wizard_options

  not_supported = {
    pageviews = true
    sessions = true
    bandwidth = true
  } # not_supported

} # urlscan