# Copyright (c) 2012 Flowerfire, Inc.  All Rights Reserved.
akamai_streaming = {

  plugin_version = "1.2"

  # 2012-05-08 - 1.0 - GMF - Initial creation
  # 2012-06-11 - 1.1 - GAS - Added media_reports snapon
  # 2013-06-25 - 1.2 - GMF - Fixed parsing of #Fields line to include final field

  info.1.manufacturer = "Akamai"
  info.1.device = "HTTP Streaming (W3C)"
  info.1.version.1 = "1.0"  # #Version line shows 1.0

  # The name of the log format
  log.format.format_label = "Akamai HTTP Streaming (W3C) Log Format"
  log.miscellaneous.log_data_type = "generic_w3c"
  log.miscellaneous.log_format_type = "media_server"

  # The log is in this format if any of the first ten lines match this regular expression
##Version: 1.0
##Fields: date time cs-ip cs-method cs-uri sc-status sc-bytes time-taken cs(Referer) cs(User-Agent) cs(Cookie)
#2012-04-27	12:05:17	12.34.56.78	GET	/dir/file.xml	200	368	0	"-"	"Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.0; Trident/5.0)"	"-"
  log.format.autodetect_regular_expression = `^#Fields: date time cs-ip cs-method cs-uri sc-status sc-bytes time-taken cs[(]Referer[)] cs[(]User-Agent[)] cs[(]Cookie[)]`

  # This handles #Fields lines, and creates log and database fields from them
  log.filter_preprocessor = `
if (matches_regular_expression(current_log_line(), '^#Fields: +([^ ].*)$')) then (
  string fields = $1;
  string fieldname;
  v.logfieldindex = 1;
  string numerical_fields = "profiles." . internal.profile_name . ".database.numerical_fields";

  # This subroutine creates a database field
  subroutine(create_database_field(string fieldname, int top, int bottom), (
    #echo("create_database_field: " . fieldname); # debug
    debug_message("create_database_field(" . fieldname . ")\n");
    string databasefieldpath = "profiles." . internal.profile_name . ".database.fields." . fieldname;
    (databasefieldpath . "") = "";
    node databasefield = databasefieldpath;
    if (top ne 0) then
      set_subnode_value(databasefield, "suppress_top", top);
    if (bottom ne 0) then
      set_subnode_value(databasefield, "suppress_bottom", bottom);
    databasefield;
  ));

  subroutine(create_log_field(string fieldname, string type, bool withindex), (
    debug_message("create_log_field(" . fieldname . "; type=" . type . ")\n");
    string logfieldpath = "profiles." . internal.profile_name . ".log.fields." . fieldname;
    (logfieldpath . "") = "";
    node logfield = logfieldpath;
    if (withindex) then (
      set_subnode_value(logfield, "index", v.logfieldindex);
      v.logfieldindex++;
    );
    set_subnode_value(logfield, "subindex", 0);
    if (type ne '') then
      set_subnode_value(logfield, "type", type);
    logfield;
  ));

  # Extract the fields on at a time
  while (matches_regular_expression(fields, '^([^ ]+) (.*)$') or
         matches_regular_expression(fields, '^([^ ]+)$')) (

    string unconverted_fieldname = $1;
    fields = $2;

    # Clean up the field name
    fieldname = '';
    for (int i = 0; i < length(unconverted_fieldname); i++) (
      string c = lowercase(substr(unconverted_fieldname, i, 1));
      if (!matches_regular_expression(c, '^[a-z0-9]$')) then
        c = '_';
      fieldname .= c;
    );
    while (matches_regular_expression(fieldname, '^(.*)_$'))
      fieldname = $1;

#echo("fieldname: " . fieldname);

    # Get the log field type
    string log_field_type = '';
    if (fieldname eq 'cs_uri') then (
      log_field_type = 'page';
    );
    else if (fieldname eq 'cs_user_agent') then (
      log_field_type = 'agent';
    );
    else if (fieldname eq 'cs_ip') then (
      log_field_type = 'host';
    );

    # Create the log field
    create_log_field(fieldname, log_field_type, true);

    # If we're creating a profile, create the database fields too.
    if (node_exists("volatile.creating_profile")) then (

      # Start the if/else chain.
      if (false) then ( );

      # Handle date by creating date_time and derived database fields
      else if (fieldname eq "date") then (
        create_database_field('date_time', 0,  0);
        create_database_field('day_of_week', 0,  0);
        create_database_field('hour_of_day', 0,  0);
      ); # if date

      else if (fieldname eq "time") then (
        create_database_field('date_time', 0,  0);
        create_database_field('day_of_week', 0,  0);
        create_database_field('hour_of_day', 0,  0);
      ); # if time

      # Create derived fields for agent
      else if (fieldname eq "cs_user_agent") then (
        create_database_field('operating_system', 0,  0);
        create_database_field('web_browser', 0,  0);
        create_database_field('spider', 0,  0);
      );

      # Create database field cs_ip and derived field for client IP 
      else if (fieldname eq "cs_ip") then (
        create_database_field('cs_ip', 0,  0);
        create_database_field('location', 0,  0);
      );

      # Create derived fields from cs_uri
      else if (fieldname eq "cs_uri") then (
      	create_database_field('cs_uri', 0,  9);
        create_database_field('file_type', 0,  0);
      );

      # Don't add a database field for numerical fields
      else if (subnode_exists(numerical_fields, fieldname)) then (
        debug_message("Not adding numerical field: " . fieldname . "\n");
      );

      # Create a normal database field
      else
        create_database_field(fieldname, 0,  0);
    ); # if creating profile

  ); # while another field

  # Don't parse the #Fields line as a data line
  'reject';

); # if #Fields

# Don't parse any other # lines as data lines
else if (starts_with(current_log_line(), '#')) then (
  'reject';
);
`

  # Get web browser, operating system, web browser, and spider information from the user-agent field.
  log.parsing_filters.derive_from_user_agent = {
    value = `get_user_agent_info(replace_all(cs_user_agent, '+', ' '));
web_browser = volatile.web_browser;
operating_system = volatile.operating_system;
spider = volatile.spider;
`
    requires_fields = {
      cs_user_agent = true
    }
  }
  
  
  
  # Log Filters
  log.filters = {

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

  } # log.filters

  database.numerical_fields = {

    events = {
      default = true
      requires_log_field = false
      entries_field = true
    } # events

    unique_client_ips = {
      label = "$lang_stats.field_labels.unique_client_ips"
      default = true
      requires_log_field = true
      log_field = "cs_ip"
      type = "unique"
    } # unique_client_ips

    sc_bytes = {
      default = true
      integer_bits = 64
      display_format_type = "bandwidth"
    }

    time_taken = {
      label = "$lang_stats.field_labels.play_duration"
      integer_bits = 64
      display_format_type = duration_milliseconds
      default = true
    } # time_taken

    average_time_taken = {
      label = "$lang_stats.field_labels.play_duration $lang_stats.field_labels.average_tag"
      integer_bits = 64
      display_format_type = duration_milliseconds
      default = false
      log_field = "time_taken"
      requires_log_field = true
      type = "int"
      aggregation_method = "average"
      average_denominator_field = "events"
    } # average_time_taken


  } # database.numerical_fields

  create_profile_wizard_options = {

    # How the reports should be grouped in the report menu
    report_groups = {
      date_time_group = ""
    } # report_groups

    snapons = {
    	
      # Attach a media_reports snapon
      media_reports = {
        snapon = "media_reports"
        name = "media_reports"
        label = "$lang_admin.snapons.media_reports.label"
        parameters = {
          user_field.parameter_value = "cs_ip"
#          have_category_field.parameter_value = false
#          category_field.parameter_value = "category"
#          domain_field.parameter_value = "top_level_domain"
          duration_field.parameter_value = "time_taken"
          stream_name_field.parameter_value = "cs_uri"
          client_ip_field.parameter_value = "cs_ip"

          # There is no stream ID in this log format
          have_stream_id_field.parameter_value = false
#          stream_id_field.parameter_value = "cs_uri"

          # Do the default concurrent connections analysis; it's based on a single duration field, which we have.
          attach_concurrent_events.parameter_value = true
  
        } # parameters

	requires_database_fields = {
	  cs_ip = true
	  time_taken = true
	  cs_uri = true
	} # requires_database_fields

      } # media_reports

    } # snapons

  } # create_profile_wizard_options

} # akamai_streaming