Module:Sensitive IP addresses/API

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

This module provides an API for information about IP addresses that Wikipedia considers sensitive. The intention is that this one API can be used for templates, Lua modules, and software using the MediaWiki Action API such as JavaScript gadgets and bots.

Usage

From templates

Templates wishing to make use of this API need to use an intermediary Lua module to parse the results of API queries. One such module, used to create a wikitable summary of sensitive IPs, exists at Module:Sensitive IP addresses/summary.

From Lua

To load this module from Lua modules, use:

local querySensitiveIPs = require('Module:Sensitive IP addresses/API').query

The query function is called with named parameters. For example:

local result = querySensitiveIPs{
	test = {'1.2.3.4', '5.6.7.8'}
}

Parameters

The following parameters are available to the query function:

  • test - an array of IP addresses and/or IP ranges to test for sensitivity. IP addresses and ranges can be IPv4 or IPv6, and ranges must be in CIDR notation.
  • entities - an array of entity IDs to get information about. An entity is a country or organization which is considered sensitive, and for which blocks should be handled with care. Entity IDs are defined in Module:Sensitive IP addresses/list along with the rest of the sensitive IP data. For example, ushr is the ID for the United States House of Representatives. If the special ID all is contained in the array, information about all entities will be included in the result.
  • format - the format to return results in. Use json to return a JSON-formatted string, and use lua to return a Lua table. If this option is not specified, a Lua table is returned by default.

Results

By default, the query function returns a Lua table, but it can return a JSON object if the format option is set to json. Whether Lua or JSON, the structure of the object returned is similar to the structure of query results from the MediaWiki Action API.

Top-level object

The top level object contains exactly one child object. If the query executed successfully, this object has a key of sensitiveips and contains the query results.

{
    "sensitiveips": {
        [query results]
    }       
}

If there were any errors when executing the query, the child of the top-level object has a key of error and contains error information. The error object has three keys: code, the error ID; info, the error message; and *, a message about where to find the API documentation. The error IDs all have a prefix of "sipa". For example:

{
    "error": {
        "code": "sipa-invalid-test-string",
        "info": "test string #1 'foo' was not a valid IP address or CIDR string",
        "*": "See https://en.wikipedia.org/wiki/Module:Sensitive_IP_addresses/API for API usage"
    }       
}

Sensitive IPs object

If the query was successful, the sensitiveips child object will be present and can contain the following objects and arrays:

  • matches - an array of IP address objects or IP range objects, where the corresponding IP address or IP range string was specified in the test query option, and where Wikipedia regards that IP address or IP range as being sensitive. If no IPs or ranges were tested for, or if no matches were found, this array will not be present in the results.
  • matched-ranges - an object with CIDR IP range strings as keys, and matched-range objects as values, where the IP range matches one of the IP addresses or IP ranges tested for with the test query option. If no IPs or ranges were tested for, or if no matches were found, this object will not be present.
  • entities - an object with entity IDs as keys, and entity objects as values. An entity is a country or organization which has IP addresses that Wikipedia considers sensitive. Entity IDs are defined in Module:Sensitive IP addresses/list; for example, ushr is the ID for the United States House of Representatives. Entities will be included in this object if an IP range belonging to them is matched by one of the IP addresses or IP ranges tested for with the test query option, or if their entity ID is specified in the entities query option.
  • entity-ids - an array of entity ID strings, in the order they are defined in Module:Sensitive IP addresses/list. The entity IDs in this array correspond one-to-one with the entity ID keys of the entities result object. This array can be useful for outputting the IDs in the same order that they were defined in the list.

IP address objects

An IP address object represents a single IPv4 or IPv6 address that matches a sensitive IP range. IP address objects contain the following fields:

  • ip - the string representation of the IP address, e.g. "1.2.3.4" or "2001:d8::ffff:ab:cdef".
  • type - the string "ip" (used to differentiate between IP address objects and IP range objects).
  • ip-version - the version of the IP protocol the address uses. This is either "IPv4" or "IPv6".
  • matches-range - the sensitive IP range that the address matches, in CIDR notation.
  • entity-id - the entity ID of the entity that owns the sensitive IP range that the address matches.

IP range objects

An IP range object represents an IPv4 or IPv6 range that overlaps with a sensitive IP range. IP range objects contain the following fields:

  • range - the CIDR string representation of the range, e.g. "1.2.3.0/24" or "2001:d8::ffff:ab:0/16".
  • type - the string "range" (used to differentiate between IP range objects and IP address objects).
  • ip-version - the version of the IP protocol the range uses. This is either "IPv4" or "IPv6".
  • matches-range - the sensitive IP range that the tested range overlaps, in CIDR notation.
  • entity-id - the entity ID of the entity that owns the sensitive IP range that the tested range overlaps.

Entity objects

An entity object represents a country or organization that has IP ranges which Wikipedia considers sensitive. Entity objects may contain the following fields:

  • id - the entity ID. This is a unique string used to identify the entity. This field is always present.
  • name - the name of the entity. This is a plain string, containing no wikitext, and is always present.
  • description - a description of the entity. This is a string, and may contain wikitext. This field is optional, and may not be present.
  • reason - the reason that the entity's IP ranges are sensitive. The possible reasons are political and technical.
  • ipv4-ranges - an array of IPv4 CIDR strings that belong to the entity, and are considered as sensitive by Wikipedia. This field is optional, and may not be present.
  • ipv6-ranges - an array of IPv6 CIDR strings that belong to the entity, and are considered as sensitive by Wikipedia. This field is optional, and may not be present.
  • notes - notes about the entity or its ranges. This field is optional, and may not be present.

Examples

Here are some examples of some queries from Lua and the results they produce.

No matches

Query:

querySensitiveIPs{
    test = {'1.2.3.4'}
}

Result:

{
  ["sensitiveips"] = {
  }
}

One match

Query:

querySensitiveIPs{
    test = {'156.33.5.76'}
}

Result:

{
  ["sensitiveips"] = {
    ["matches"] = {
      {
        ["type"] = "ip",
        ["ip"] = "156.33.5.76",
        ["ip-version"] = "IPv4",
        ["matches-range"] = "156.33.0.0/16",
        ["entity-id"] = "ussenate",
      },
    },
    ["matched-ranges"] = {
      ["156.33.0.0/16"] = {
        ["range"] = "156.33.0.0/16",
        ["ip-version"] = "IPv4",
        ["entity-id"] = "ussenate",
      },
    },
    ["entities"] = {
      ["ussenate"] = {
        ["id"] = "ussenate",
        ["name"] = "United States Senate",
        ["description"] = "the [[United States Senate]]",
        ["reason"] = "political",
        ["ipv4Ranges"] = {
          "156.33.0.0/16",
        },
        ["ipv6Ranges"] = {
          "2620:0:8a0::/48",
          "2600:803:618::/48",
        }
      },
    },
    ["entity-ids"] = {
      "ussenate",
    },
  },
}

One match, JSON output

Query: Query:

querySensitiveIPs{
    format = 'json',
    test = {'156.33.5.76'}
}

Result:

{
   "sensitiveips":{
      "matches":[
         {
            "type": "ip",
            "ip": "156.33.5.76",
            "ip-version": "IPv4",
            "matches-range": "156.33.0.0/16",
            "entity-id": "ussenate"
         }
      ],
      "matched-ranges": {
         "156.33.0.0/16": {
            "range": "156.33.0.0/16",
            "ip-version": "IPv4",
            "entity-id": "ussenate"
         }
      },
      "entities": {
         "ussenate": {
            "id": "ussenate",
            "name": "United States Senate",
            "description": "the [[United States Senate]]",
            "reason": "political",
            "ipv6Ranges": [
               "2620:0:8a0::/48",
               "2600:803:618::/48"
            ],
            "ipv4Ranges": [
               "156.33.0.0/16"
            ]
         }
      },
      "entity-ids": [
         "ussenate"
      ]
   }
}

Entity IDs

querySensitiveIPs{
    format = 'json',
    entities = {'usdhs', 'usdoj'}
}

Result:

{
   "sensitiveips": {
      "entities": {
         "usdoj": {
            "id": "usdoj",
            "name": "United States Department of Justice",
            "description": "the [[United States Department of Justice]]",
            "reason": "political",
            "ipv4Ranges": [
               "149.101.0.0/16"
            ]
         },
         "usdhs": {
            "id": "usdhs",
            "name": "United States Department of Homeland Security",
            "description": "the [[United States Department of Homeland Security]]",
            "reason": "political",
            "ipv4Ranges": [
               "65.165.132.0/24",
               "204.248.24.0/24",
               "216.81.80.0/20"
            ]
         }
      },
      "entity-ids": [
         "usdoj",
         "usdhs"
      ]
   }
}

Invalid IP error

Query:

querySensitiveIPs{
    test = {'foo'}
}

Result:

{
  ["error"] = {
    ["code"] = "sipa-invalid-test-string",
    ["info"] = "test string #1 'foo' was not a valid IP address or CIDR string"
    ["*"] = "See https://en.wikipedia.org/wiki/Module:Sensitive_IP_addresses/API for API usage",
  }
}



-- This module provides functions for handling sensitive IP addresses.

-- Load modules
local mIP = require('Module:IP')
local IPAddress = mIP.IPAddress
local Subnet = mIP.Subnet
local IPv4Collection = mIP.IPv4Collection
local IPv6Collection = mIP.IPv6Collection

-- Lazily load the jf-JSON module
local JSON

-------------------------------------------------------------------------------
-- Helper functions
-------------------------------------------------------------------------------

local function deepCopy(val)
	-- Make a deep copy of a value, but don't worry about self-references or
	-- metatables as mw.clone does. If a table in val has a self-reference,
	-- you will get an infinite loop, so don't do that.
	if type(val) == 'table' then
		local ret = {}
		for k, v in pairs(val) do
			ret[k] = deepCopy(v)
		end
		return ret
	else
		return val
	end
end

local function deepCopyInto(source, dest)
	-- Do a deep copy of a source table into a destination table, ignoring
	-- self-references and metatables. If a table in source has a self-reference
	-- you will get an infinite loop.
	for k, v in pairs(source) do
		if type(v) == 'table' then
			dest[k] = {}
			deepCopyInto(v, dest[k])
		else
			dest[k] = v
		end
	end
end

local function removeDuplicates(t)
	-- Return a copy of an array with duplicate values removed.
	local keys, ret = {}, {}
	for i, v in ipairs(t) do
		if not keys[v] then
			table.insert(ret, v)
			keys[v] = true
		end
	end
	return ret
end

-------------------------------------------------------------------------------
-- SensitiveEntity class
-- A country or organization for which blocks must be handled with care.
-- Media organizations may inspect block messages for IP addresses and ranges
-- belonging to these entities and those messages may end up in the press.
-------------------------------------------------------------------------------

local SensitiveEntity = {}
SensitiveEntity.__index = SensitiveEntity

SensitiveEntity.reasons = {
	-- The reasons that an entity may be sensitive. Used to verify data in
	-- Module:Sensitive IP addresses/list.
	political = true,
	technical = true,
}

do
	-- Private methods
	local function addRanges(self, key, collectionConstructor, ranges)
		if ranges and ranges[1] then
			self[key] = collectionConstructor()
			for i, range in ipairs(ranges) do
				self[key]:addSubnet(Subnet.new(range))
			end
		end
	end

	-- Constructor
	function SensitiveEntity.new(data)
		local self = setmetatable({}, SensitiveEntity)

		-- Set data
		self.data = data
		addRanges(self, 'v4Collection', IPv4Collection.new, data.ipv4Ranges)
		addRanges(self, 'v6Collection', IPv6Collection.new, data.ipv6Ranges)

		return self
	end
end

function SensitiveEntity:matchesIPOrRange(str)
	-- Returns true, matchObj, queryObj if there is a match for the IP address
	-- string or CIDR range str in the sensitive entity. Returns false
	-- otherwise. matchObj is the Subnet object that was matched, and queryObj
	-- is the IPAddress or Subnet object corresponding to the input string.

	-- Get the IPAddress or Subnet object for str
	local isIP, isSubnet, obj
	isIP, obj = pcall(IPAddress.new, str)
	if isIP and not obj then
		isIP = false
	end

	if not isIP then
		isSubnet, obj = pcall(Subnet.new, str)
		if not isSubnet or not obj then
			error(string.format(
				"'%s' is not a valid IP address or CIDR string",
				str
			), 2)
		end
	end

	-- Try matching the object to the appropriate collection
	local function isInCollection(collection, obj, isIP)
		if isIP then
			if collection then
				local isMatch, matchObj = collection:containsIP(obj)
				return isMatch, matchObj, obj
			else
				return false
			end
		else
			if collection then
				local isMatch, matchObj = collection:overlapsSubnet(obj)
				return isMatch, matchObj, obj
			else
				return false
			end
		end
	end

	if obj:isIPv4() then
		return isInCollection(self.v4Collection, obj, isIP)
	else
		return isInCollection(self.v6Collection, obj, isIP)
	end
end

-------------------------------------------------------------------------------
-- Sensitive IP API
-------------------------------------------------------------------------------

-- This API is used by external tools and gadgets, so it should be kept
-- backwards-compatible. Clients query the API with a query table, and the
-- API returns a response table. The response table is available as a Lua table
-- for other Lua modules, and as JSON for external clients.

-- Example query tables:
--
-- Query IP addresses and ranges:
-- {
-- 	test = {'1.2.3.4', '4.5.6.0/24', '2001:db8::ff00:12:3456', '2001:db8::ff00:12:0/112'},
-- }
--
-- Query specific entities:
-- {
-- 	entities = {'ussenate', 'ushr'}
-- }
--
-- Query all entities:
-- {
-- 	entities = {'all'}
-- }
--
-- Query all entities and format the result as a JSON string:
-- {
-- 	entities = {'all'},
--  format = 'json'
-- }
--
-- Combined query:
-- {
-- 	test = {'1.2.3.4', '4.5.6.0/24', '2001:db8::ff00:12:3456', '2001:db8::ff00:12:0/112'},
-- 	entities = {'ussenate', 'ushr'}
-- }

-- Example response:
--
-- {
--     sensitiveips = {
--         matches = {
--             {
--                 ip = '1.2.3.4',
--                 type = 'ip',
--                 ['ip-version'] = 'IPv4',
--                 ['matches-range'] = '1.2.3.0/24',
--                 ['entity-id'] = 'entityid'
--             },
--             {
--                 range = '4.5.6.0/24',
--                 type = 'range',
--                 ['ip-version'] = 'IPv4',
--                 ['matches-range'] = '4.5.0.0/16',
--                 ['entity-id'] = 'entityid'
--             }
--         },
--         ['matched-ranges'] = {
--             ['1.2.3.0/24'] = {
--                 range = '1.2.3.0/24',
--                 ['ip-version'] = 'IPv4',
--                 ['entity-id'] = 'entityid'
--             },
--             ['4.5.0.0/16'] = {
--                 range = '4.5.0.0/16',
--                 ['ip-version'] = 'IPv4',
--                 ['entity-id'] = 'entityid'
--             }
--         },
--         entities = {
--             ['entityid'] = {
--                 id = 'entityid',
--                 name = 'The entity name',
--                 description = 'A description of the entity',
--                 ['ipv4-ranges'] = {
--                     '1.2.3.0/24',
--                     '4.5.0.0/16'
--                     '6.7.0.0/16'
--                 },
--                 ['ipv6-ranges'] = {
--                     '2001:db8::ff00:12:0/112'
--                 },
--                 notes = 'Notes about the entity or its ranges'
--             }
--         }
--         ['entity-ids'] = {
--             'entityid'
--         }
--     }
-- }
--
-- Response with errors:
--
-- {
--     error = {
--         code = 'example-error',
--         info = 'There was an error',
--         ['*'] = 'See https://en.wikipedia.org/wiki/Module:Sensitive_IP_addresses for API usage'
--     }
-- }

local function query(options)
	-- Make entity objects
	local entities, entityIndexes = {}, {}
	local data = mw.loadData('Module:Sensitive IP addresses/list')
	for i, entityData in ipairs(data) do
		entities[entityData.id] = SensitiveEntity.new(entityData)
		entityIndexes[entityData.id] = i -- Keep track of the original order
	end

	local function makeError(code, info, format)
		local ret = {['error'] = {
			code = code,
			info = info,
			['*'] = 'See https://en.wikipedia.org/wiki/Module:Sensitive_IP_addresses/API for API usage',
		}}
		if format == 'json' then
			return mw.text.jsonEncode(ret)
		else
			return ret
		end
	end

	-- Construct result
	local result = {
		matches = {},
		['matched-ranges'] = {},
		entities = {},
		['entity-ids'] = {}
	}

	if type(options) ~= 'table' then
		return makeError(
			'sipa-options-type-error',
			string.format(
				"type error in argument #1 of 'query' (expected table, received %s)",
				type(options)
			)
		)
	elseif not options.test and not options.entities then
		return makeError(
			'sipa-blank-options',
			"the options table didn't contain a 'test' or an 'entities' key",
			options.format
		)
	end

	if options.test then
		if type(options.test) ~= 'table' then
			return makeError(
				'sipa-test-type-error',
				string.format(
					"'test' options key was type %s (expected table)",
					type(options.test)
				),
				options.format
			)
		end

		for i, testString in ipairs(options.test) do
			if type(testString) ~= 'string' then
				return makeError(
					'sipa-test-string-type-error',
					string.format(
						"type error in item #%d in the 'test' array (expected string, received %s)",
						i,
						type(testString)
					),
					options.format
				)
			end

			for k, entity in pairs(entities) do
				-- Try to match the range with the current sensitive entity.
				local success, isMatch, matchObj, queryObj = pcall(
					entity.matchesIPOrRange,
					entity,
					testString
				)
				if not success then
					-- The string was invalid.
					return makeError(
						'sipa-invalid-test-string',
						string.format(
							"test string #%d '%s' was not a valid IP address or CIDR string",
							i,
							testString
						),
						options.format
					)
				end
				if isMatch then
					-- The string was a sensitive IP address or subnet.

					-- Add match data
					local match = {}
					-- Quick and dirty hack to find if queryObj is an IPAddress object.
					local isIP = queryObj.getNextIP ~= nil and queryObj.isInSubnet ~= nil
					if isIP then
						match.type = 'ip'
						match.ip = tostring(queryObj)
					else
						match.type = 'range'
						match.range = tostring(queryObj)
					end
					match['ip-version'] = queryObj:getVersion()
					match['matches-range'] = matchObj:getCIDR()
					match['entity-id'] = entity.data.id
					table.insert(result.matches, match)

					-- Add the matched range data.
					result['matched-ranges'][match['matches-range']] = {
						range = match['matches-range'],
						['ip-version'] = match['ip-version'],
						['entity-id'] = match['entity-id'],
					}

					-- Add the entity data for the entity we matched.
					result.entities[match['entity-id']] = deepCopy(
						entities[match['entity-id']].data
					)

					-- Add the entity ID for the entity we matched.
					table.insert(result['entity-ids'], match['entity-id'])
				end
			end
		end
	end

	-- Add entity data requested explicitly.
	if options.entities then
		if type(options.entities) ~= 'table' then
			return makeError(
				'sipa-entities-type-error',
				string.format(
					"'entities' options key was type %s (expected table)",
					type(options.test)
				),
				options.format
			)
		end

		-- Check the type of all the entity strings, and check if 'all' has
		-- been specified.
		local isAll = false
		for i, entityString in ipairs(options.entities) do
			if type(entityString) ~= 'string' then
				return makeError(
					'sipa-entity-string-type-error',
					string.format(
						"type error in item #%d in the 'entities' array (expected string, received %s)",
						i,
						type(entityString)
					),
					options.format
				)
			end
			if entityString == 'all' then
				isAll = true
			end
		end

		if isAll then
			-- Add all the entity data.
			-- As the final result will contain all the entity data, we can
			-- just create the entities and entity-ids subtables from scratch
			-- without worrying about what any existing values might be.
			result.entities = {}
			result['entity-ids'] = {}
			for i, entityData in ipairs(data) do
				result.entities[entityData.id] = deepCopy(entityData)
				result['entity-ids'][i] = entityData.id
			end
		else
			-- Add data for the entities specified.
			-- Insert the entity and entity-id subtables if they aren't already
			-- present.
			for i, entityString in ipairs(options.entities) do
				if entities[entityString] then
					result.entities[entityString] = deepCopy(
						entities[entityString].data
					)
					table.insert(result['entity-ids'], entityString)
				end
			end
			result['entity-ids'] = removeDuplicates(result['entity-ids'])
			table.sort(result['entity-ids'], function(s1, s2)
				return entityIndexes[s1] < entityIndexes[s2]
			end)
		end
	end

	-- Add any missing reason fields from entities.
	for id, entityData in pairs(result.entities) do
		entityData.reason = entityData.reason or 'political'
	end

	-- Wrap the result in an outer layer like the MediaWiki Action API does.
	result = {sensitiveips = result}

	if options.format == 'json' then
		-- Load jf-JSON
		JSON = JSON or require('Module:jf-JSON')
		JSON.strictTypes = true -- Necessary for correct blank-object encoding
		-- Decode a skeleton result JSON string. This ensures that blank objects
		-- are re-encoded as blank objects and not as blank arrays.
		local jsonResult = JSON:decode([[{"sensitiveips": {
			"matches": [],
			"matched-ranges": {},
			"entities": {},
			"entity-ids": []
		}}]])
		for i, key in ipairs{'matches', 'matched-ranges', 'entities', 'entity-ids'} do
			deepCopyInto(result.sensitiveips[key], jsonResult.sensitiveips[key])
		end
		return JSON:encode(jsonResult)
	elseif options.format == nil or options.format == 'lua' then
		return result
	elseif type(options.format) ~= 'string' then
		return makeError(
			'sipa-format-type-error',
			string.format(
				"'format' options key was type %s (expected string or nil)",
				type(options.format)
			)
		)
	else
		return makeError(
			'sipa-invalid-format',
			string.format(
				"invalid format '%s' (expected 'json' or 'lua')",
				type(options.format)
			)
		)
	end
end

--------------------------------------------------------------------------------
-- Exports
--------------------------------------------------------------------------------

local p = {}

function p._isValidSensitivityReason(s)
	-- Return true if s is a valid sensitivity reason; otherwise return false.
	return s ~= nil and SensitiveEntity.reasons[s] ~= nil
end

function p._getSensitivityReasons(separator, conjunction)
	-- Return an string of valid sensitivity reasons, ordered alphabetically.
	-- The reasons are separated by an optional separator; if conjunction is
	-- specified it is used instead of the last separator, as in
	-- mw.text.listToText.

	-- Get an array of valid sensitivity reasons.
	local reasons = {}
	for reason in pairs(SensitiveEntity.reasons) do
		reasons[#reasons + 1] = reason
	end
	table.sort(reasons)

	-- Convert arguments if we are being called from wikitext.
	if type(separator) == 'table' and type(separator.getParent) == 'function' then
		-- separator is a frame object
		local frame = separator
		separator = frame.args[1]
		conjunction = frame.args[2]
	end

	-- Return a formatted string
	return mw.text.listToText(reasons, separator, conjunction)
end

-- Export the API query function
p.query = query

return p