From a9aa9e93ab46f067bfe912aae3612a17f2743720 Mon Sep 17 00:00:00 2001 From: kientzle Date: Tue, 18 Mar 2008 18:24:32 +0000 Subject: Move the "freebase.api" module into a "freebase-api" directory so that we can provide a couple of different parallel implementations. git-svn-id: http://freebase-python.googlecode.com/svn/trunk@34 5914aa95-5b3a-0410-a3b5-7b719e7fe9b2 --- freebase/__init__.py | 0 freebase/api/__init__.py | 4 - freebase/api/httplib2cookie.py | 149 ------------ freebase/api/mqlkey.py | 129 ---------- freebase/api/session.py | 524 ----------------------------------------- freebase/rison.py | 308 ------------------------ freebase/uritemplate.py | 143 ----------- 7 files changed, 1257 deletions(-) delete mode 100755 freebase/__init__.py delete mode 100755 freebase/api/__init__.py delete mode 100755 freebase/api/httplib2cookie.py delete mode 100755 freebase/api/mqlkey.py delete mode 100644 freebase/api/session.py delete mode 100755 freebase/rison.py delete mode 100755 freebase/uritemplate.py (limited to 'freebase') diff --git a/freebase/__init__.py b/freebase/__init__.py deleted file mode 100755 index e69de29..0000000 diff --git a/freebase/api/__init__.py b/freebase/api/__init__.py deleted file mode 100755 index c3adebc..0000000 --- a/freebase/api/__init__.py +++ /dev/null @@ -1,4 +0,0 @@ - -from session import HTTPMetawebSession, MetawebError, attrdict - -from mqlkey import quotekey, unquotekey diff --git a/freebase/api/httplib2cookie.py b/freebase/api/httplib2cookie.py deleted file mode 100755 index fee0ec0..0000000 --- a/freebase/api/httplib2cookie.py +++ /dev/null @@ -1,149 +0,0 @@ -# ======================================================================== -# Copyright (c) 2007, Metaweb Technologies, Inc. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above -# copyright notice, this list of conditions and the following -# disclaimer in the documentation and/or other materials provided -# with the distribution. -# -# THIS SOFTWARE IS PROVIDED BY METAWEB TECHNOLOGIES AND CONTRIBUTORS -# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS -# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL METAWEB -# TECHNOLOGIES OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, -# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, -# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; -# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER -# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT -# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN -# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -# POSSIBILITY OF SUCH DAMAGE. -# ======================================================================== - -# -# -# httplib2cookie.py allows you to use python's standard -# CookieJar class with httplib2. -# -# - -import re -import httplib2 -from httplib2 import Http - -import mimetools -import urllib -import urllib2 -import cookielib - -class DummyRequest(object): - """Simulated urllib2.Request object for httplib2 - - implements only what's necessary for cookielib.CookieJar to work - """ - def __init__(self, url, headers=None): - self.url = url - self.headers = headers - self.origin_req_host = cookielib.request_host(self) - self.type, r = urllib.splittype(url) - self.host, r = urllib.splithost(r) - if self.host: - self.host = urllib.unquote(self.host) - - def get_full_url(self): - return self.url - - def get_origin_req_host(self): - # TODO to match urllib2 this should be different for redirects - return self.origin_req_host - - def get_type(self): - return self.type - - def get_host(self): - return self.host - - def get_header(self, key, default=None): - return self.headers.get(key.lower(), default) - - def has_header(self, key): - return key in self.headers - - def add_unredirected_header(self, key, val): - # TODO this header should not be sent on redirect - self.headers[key.lower()] = val - - def is_unverifiable(self): - # TODO to match urllib2, this should be set to True when the - # request is the result of a redirect - return False - - -class DummyResponse(object): - """Simulated urllib2.Request object for httplib2 - - implements only what's necessary for cookielib.CookieJar to work - """ - def __init__(self, response): - self.response = response - - def info(self): - return DummyMessage(self.response) - - -class DummyMessage(object): - """Simulated mimetools.Message object for httplib2 - - implements only what's necessary for cookielib.CookieJar to work - """ - def __init__(self, response): - self.response = response - - def getheaders(self, k): - k = k.lower() - v = self.response.get(k.lower(), None) - if k not in self.response: - return [] - #return self.response[k].split(re.compile(',\\s*')) - - # httplib2 joins multiple values for the same header - # using ','. but the netscape cookie format uses ',' - # as part of the expires= date format. so we have - # to split carefully here - header.split(',') won't do it. - HEADERVAL= re.compile(r'\s*(([^,]|(,\s*\d))+)') - return [h[0] for h in HEADERVAL.findall(self.response[k])] - -class CookiefulHttp(Http): - """Subclass of httplib2.Http that keeps cookie state - - constructor takes an optional cookiejar=cookielib.CookieJar - - currently this does not handle redirects completely correctly: - if the server redirects to a different host the original - cookies will still be sent to that host. - """ - def __init__(self, cookiejar=None, **kws): - # note that httplib2.Http is not a new-style-class - Http.__init__(self, **kws) - if cookiejar is None: - cookiejar = cookielib.CookieJar() - self.cookiejar = cookiejar - - def request(self, uri, **kws): - headers = kws.pop('headers', None) - req = DummyRequest(uri, headers) - self.cookiejar.add_cookie_header(req) - headers = req.headers - - (r, body) = Http.request(self, uri, headers=headers, **kws) - - resp = DummyResponse(r) - self.cookiejar.extract_cookies(resp, req) - - return (r, body) diff --git a/freebase/api/mqlkey.py b/freebase/api/mqlkey.py deleted file mode 100755 index 98b7d83..0000000 --- a/freebase/api/mqlkey.py +++ /dev/null @@ -1,129 +0,0 @@ -# ======================================================================== -# Copyright (c) 2007, Metaweb Technologies, Inc. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above -# copyright notice, this list of conditions and the following -# disclaimer in the documentation and/or other materials provided -# with the distribution. -# -# THIS SOFTWARE IS PROVIDED BY METAWEB TECHNOLOGIES AND CONTRIBUTORS -# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS -# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL METAWEB -# TECHNOLOGIES OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, -# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, -# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; -# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER -# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT -# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN -# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -# POSSIBILITY OF SUCH DAMAGE. -# ======================================================================== - -import string -import re - -def quotekey(ustr): - """ - quote a unicode string to turn it into a valid namespace key - - """ - valid_always = string.ascii_letters + string.digits - valid_interior_only = valid_always + '_-' - - if isinstance(ustr, str): - s = unicode(ustr,'utf-8') - elif isinstance(ustr, unicode): - s = ustr - else: - raise ValueError, 'quotekey() expects utf-8 string or unicode' - - output = [] - if s[0] in valid_always: - output.append(s[0]) - else: - output.append('$%04X' % ord(s[0])) - - for c in s[1:-1]: - if c in valid_interior_only: - output.append(c) - else: - output.append('$%04X' % ord(c)) - - if len(s) > 1: - if s[-1] in valid_always: - output.append(s[-1]) - else: - output.append('$%04X' % ord(s[-1])) - - return str(''.join(output)) - - -def unquotekey(key, encoding=None): - """ - unquote a namespace key and turn it into a unicode string - """ - - valid_always = string.ascii_letters + string.digits - - output = [] - i = 0 - while i < len(key): - if key[i] in valid_always: - output.append(key[i]) - i += 1 - elif key[i] in '_-' and i != 0 and i != len(key): - output.append(key[i]) - i += 1 - elif key[i] == '$' and i+4 < len(key): - # may raise ValueError if there are invalid characters - output.append(unichr(int(key[i+1:i+5],16))) - i += 5 - else: - raise ValueError, "unquote key saw invalid character '%s' at position %d" % (key[i], i) - - ustr = u''.join(output) - - if encoding is None: - return ustr - - return ustr.encode(encoding) - - - -def id_to_urlid(id): - """ - convert a mql id to an id suitable for embedding in a url path. - """ - - # XXX shouldn't be in metaweb.api! - from mw.formats.http import urlencode_pathseg - - segs = id.split('/') - - assert isinstance(id, str) and id != '', 'bad id "%s"' % id - - if id[0] == '~': - assert len(segs) == 1 - # assume valid, should check - return id - - if id[0] == '#': - assert len(segs) == 1 - # assume valid, should check - return '%23' + id[1:] - - if id[0] != '/': - raise ValueError, 'unknown id format %s' % id - - # ok, we have a slash-path - # requote components as keys and rejoin. - # urlids do not have leading slashes!!! - return '/'.join(urlencode_pathseg(unquotekey(seg)) for seg in segs[1:]) - diff --git a/freebase/api/session.py b/freebase/api/session.py deleted file mode 100644 index fc32676..0000000 --- a/freebase/api/session.py +++ /dev/null @@ -1,524 +0,0 @@ -# ======================================================================== -# Copyright (c) 2007, Metaweb Technologies, Inc. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above -# copyright notice, this list of conditions and the following -# disclaimer in the documentation and/or other materials provided -# with the distribution. -# -# THIS SOFTWARE IS PROVIDED BY METAWEB TECHNOLOGIES AND CONTRIBUTORS -# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS -# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL METAWEB -# TECHNOLOGIES OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, -# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, -# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; -# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER -# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT -# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN -# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -# POSSIBILITY OF SUCH DAMAGE. -# ======================================================================== - -# -# declarations for external metaweb api. -# -# -# from metaweb.api import HTTPMetawebSession -# -# mss = HTTPMetawebSession('sandbox.freebase.com') -# print mss.mqlread([dict(name=None, type='/type/type')]) -# -# -# - -__all__ = ['MetawebError', 'MetawebSession', 'HTTPMetawebSession', 'attrdict'] -__version__ = '0.1' - -import os, sys, re -import urllib2 -import cookielib -import simplejson -from urllib import quote as urlquote -import pprint -import socket -import logging - -try: - import httplib2 - from httplib2cookie import CookiefulHttp -except ImportError: - httplib2 = None - CookiefulHttp = None - print ('freebase.api: you can install httplib2 for better performance') - -import simplejson.encoder -# remove whitespace from json encoded output -simplejson.JSONEncoder.item_separator = ',' -simplejson.JSONEncoder.key_separator = ':' -# don't escape slashes, we're not pasting into script tags here. -if simplejson.encoder.ESCAPE_DCT.get('/', None) == r'\/': - simplejson.encoder.ESCAPE_DCT['/'] = '/' - -def urlencode_weak(s): - return urlquote(s, safe=',/:$') - - -# from http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/361668 -class attrdict(dict): - """A dict whose items can also be accessed as member variables. - - >>> d = attrdict(a=1, b=2) - >>> d['c'] = 3 - >>> print d.a, d.b, d.c - 1 2 3 - >>> d.b = 10 - >>> print d['b'] - 10 - - # but be careful, it's easy to hide methods - >>> print d.get('c') - 3 - >>> d['get'] = 4 - >>> print d.get('a') - Traceback (most recent call last): - TypeError: 'int' object is not callable - """ - def __init__(self, *args, **kwargs): - dict.__init__(self, *args, **kwargs) - self.__dict__ = self - - - -# TODO expose the common parts of the result envelope -class MetawebError(Exception): - """ - an error report from the metaweb service. - """ - pass - - - -# TODO right now this is a completely unnecessary superclass. -# is there enough common behavior between session types -# to justify it? -class MetawebSession(object): - """ - MetawebSession is the base class for MetawebSession, subclassed for - different connection types. Only http is available externally. - - This is more of an interface than a class - """ - - # interface definition here... - - -# from httplib2 -NORMALIZE_SPACE = re.compile(r'(?:\r\n)?[ \t]+') -def _normalize_headers(headers): - return dict([ (key.lower(), NORMALIZE_SPACE.sub(value, ' ').strip()) for (key, value) in headers.iteritems()]) - - -class HTTPMetawebSession(MetawebSession): - """ - a MetawebSession is a request/response queue. - - this version uses the HTTP api, and is synchronous. - """ - # share cookies across sessions, so that different sessions can - # see each other's writes immediately. - _default_cookiejar = cookielib.CookieJar() - - def __init__(self, service_url, username=None, password=None, prev_session=None, cookiejar=None, cookiefile=None): - """ - create a new MetawebSession for interacting with the Metaweb. - - a new session will inherit state from prev_session if present, - """ - super(HTTPMetawebSession, self).__init__() - - self.log = logging.getLogger() - - assert not service_url.endswith('/') - if not '/' in service_url: # plain host:port - service_url = 'http://' + service_url - - self.service_url = service_url - - self.username = username - self.password = password - - self.tid = None - - if prev_session: - self.service_url = prev.service_url - - if cookiefile is not None: - cookiejar = self.open_cookie_file(cookiefile) - - if cookiejar is not None: - self.cookiejar = cookiejar - elif prev_session: - self.cookiejar = prev_session.cookiejar - else: - self.cookiejar = self._default_cookiejar - - if CookiefulHttp is not None: - self.httpclient = CookiefulHttp(cookiejar=self.cookiejar) - else: - cookiespy = urllib2.HTTPCookieProcessor(self.cookiejar) - self.opener = urllib2.build_opener(cookiespy) - - - def open_cookie_file(self, cookiefile=None): - if cookiefile is None or cookiefile == '': - if os.environ.has_key('HOME'): - cookiefile = os.path.join(os.environ['HOME'], '.pyfreebase/cookiejar') - else: - raise MetawebError("no cookiefile specified and no $HOME/.pyfreebase directory" % cookiefile) - - cookiejar = cookielib.LWPCookieJar(cookiefile) - if os.path.exists(cookiefile): - cookiejar.load(ignore_discard=True) - - return cookiejar - - - def _httpreq(self, service_path, method='GET', body=None, form=None, - headers=None): - """ - make an http request to the service. - - form arguments are encoded in the url, even for POST, if a non-form - content-type is given for the body. - - returns a pair (resp, body) - - resp is the response object and may be different depending - on whether urllib2 or httplib2 is in use? - """ - - if method == 'POST': - assert body is not None or form is not None - elif method == 'GET': - assert body is None - else: - assert 0, 'unknown method %s' % method - - url = self.service_url + service_path - - if headers is None: - headers = {} - else: - headers = _normalize_headers(headers) - - # XXX This is a lousy way to parse Content-Type, where is - # the library? - ct = headers.get('content-type', None) - if ct is not None: - ct = ct.split(';')[0] - - if body is not None: - # if body is provided, content-type had better be too - assert ct is not None - - if form is not None: - qstr = '&'.join(['%s=%s' % (urlencode_weak(k), urlencode_weak(v)) - for k,v in form.items()]) - if method == 'POST': - # put the args on the url if we're putting something else - # in the body. this is used to add args to raw uploads. - if body is not None: - url += '?' + qstr - else: - if ct is None: - # XXX encoding and stuff - ct = 'application/x-www-form-urlencoded' - headers['content-type'] = ct - - if ct == 'multipart/form-encoded': - # XXX fixme - raise NotImplementedError - elif ct == 'application/x-www-form-urlencoded': - body = qstr - else: - # for all methods other than POST, use the url - url += '?' + qstr - - - # assure the service that this isn't a CSRF form submission - headers['x-metaweb-request'] = 'Python' - - if 'user-agent' not in headers: - headers['user-agent'] = 'python freebase.api-%s' % __version__ - - #if self.tid is not None: - # headers['x-metaweb-tid'] = self.tid - - ####### DEBUG MESSAGE - should check log level before generating - if form is None: - formstr = '' - else: - formstr = 'FORM:\n ' + '\n '.join(['%s=%s' % (k,v) - for k,v in form.items()]) - if headers is None: - headerstr = '' - else: - headerstr = 'HEADERS:\n ' + '\n '.join([('%s: %s' % (k,v)) - for k,v in headers.items()]) - self.log.debug('%s %s%s%s', method, url, formstr, headerstr) - ####### - - if CookiefulHttp is not None: - return self._httplib2_request(url, method, body, headers) - else: - return self._urllib2_request(url, method, body, headers) - - - def _raise_service_error(self, status, ctype, body): - is_jsbody = (e.info().type.endswith('javascript') - or e.info().type.endswith('json')) - if str(status) == '400' and is_jsbody: - r = self._loadjson(body) - msg = r.messages[0] - raise MetawebError(u'%s %s %r' % (msg.get('code',''), msg.message, msg.info)) - - raise MetawebError, 'request failed: %s: %r %r' % (url, str(e), body) - - def _urllib2_request(self, url, method, body, headers): - req = urllib2.Request(url, body, headers) - - try: - resp = self.opener.open(req) - - except socket.error, e: - self.log.error('SOCKET FAILURE: %s', e.fp.read()) - raise MetawebError, 'failed contacting %s: %s' % (url, str(e)) - - except urllib2.HTTPError, e: - _raise_service_error(e.code, e.info().type, e.fp.read()) - - for header in resp.info().headers: - self.log.debug('HTTP HEADER %s', header) - name, value = re.split("[:\n\r]", header, 1) - if name.lower() == 'x-metaweb-tid': - self.tid = value.strip() - - return (resp, resp.read()) - - def _httplib2_request(self, url, method, body, headers): - try: - resp, content = self.httpclient.request(url, method=method, - body=body, headers=headers) - except socket.error, e: - self.log.error('SOCKET FAILURE: %s', e.fp.read()) - raise MetawebError, 'failed contacting %s: %s' % (url, str(e)) - - except httplib2.HttpLib2ErrorWithResponse, e: - self._raise_service_error(resp.status, resp['content-type'], content) - except httplib2.HttpLib2Error, e: - raise MetawebError(u'HTTP error: %s' % (e,)) - - #tid = resp.get('x-metaweb-tid', None) - - return (resp, content) - - - def _httpreq_json(self, *args, **kws): - resp, body = self._httpreq(*args, **kws) - return self._loadjson(body) - - def _loadjson(self, json): - # TODO really this should be accomplished by hooking - # simplejson to create attrdicts instead of dicts. - def struct2attrdict(st): - """ - copy a json structure, turning all dicts into attrdicts. - - copying descends instances of dict and list, including subclasses. - """ - if isinstance(st, dict): - return attrdict([(k,struct2attrdict(v)) for k,v in st.items()]) - if isinstance(st, list): - return [struct2attrdict(li) for li in st] - return st - - if json == '': - self.log.error('the empty string is not valid json') - raise MetawebError('the empty string is not valid json') - - try: - r = simplejson.loads(json) - except ValueError, e: - self.log.error('error parsing json string %r' % json) - raise MetawebError, 'error parsing JSON string: %s' % e - - return struct2attrdict(r) - - def _check_mqlerror(self, r): - if r.code != '/api/status/ok': - for msg in r.messages: - self.log.error('mql error: %s %s %r' % (msg.code, msg.message, msg.get('query', None))) - raise MetawebError, 'query failed: %s %r' % (r.messages[0].code, r.messages[0].get('query', None)) - - def _mqlresult(self, r): - self._check_mqlerror(r) - - # should check log level to avoid redundant simplejson.dumps - rstr = simplejson.dumps(r.result, indent=2) - if rstr[0] == '{': - rstr = rstr[1:-2] - self.log.info('result: %s', rstr) - - return r.result - - - - def login(self): - """sign in to the service""" - - assert self.username is not None - assert self.password is not None - - self.log.debug('LOGIN USERNAME: %s', self.username) - - try: - r = self._httpreq_json('/api/account/login', 'POST', - form=dict(username=self.username, - password=self.password)) - except urllib2.HTTPError, e: - raise MetawebError("login error: %s", e) - - if r.code != '/api/status/ok': - raise MetawebError(u'%s %r' % (r.get('code',''), r.messages)) - - self.log.debug('LOGIN RESP: %r', r) - self.log.debug('LOGIN COOKIES: %s', self.cookiejar) - - - def mqlreaditer(self, sq): - """read a structure query""" - - cursor = True - - while 1: - subq = dict(query=[sq], cursor=cursor, escape=False) - qstr = simplejson.dumps(subq) - - service = '/api/service/mqlread' - - r = self._httpreq_json(service, form=dict(query=qstr)) - - for item in self._mqlresult(r): - yield item - - if r['cursor']: - cursor = r['cursor'] - self.log.info('CONTINUING with %s', cursor) - else: - return - - def mqlread(self, sq): - """read a structure query""" - subq = dict(query=sq, escape=False) - if isinstance(sq, list): - subq['cursor'] = True - - service = '/api/service/mqlread' - - # should check log level to avoid redundant simplejson.dumps - self.log.info('%s: %s', - service, - simplejson.dumps(sq, indent=2)[1:-2]) - - qstr = simplejson.dumps(subq) - r = self._httpreq_json(service, form=dict(query=qstr)) - - return self._mqlresult(r) - - def trans(self, guid): - """translate blob from guid """ - url = '/api/trans/raw' + urlquote(guid) - - self.log.info(url) - - resp, body = self._httpreq(url) - - self.log.info('%d bytes' % len(body)) - - return body - - def mqlwrite(self, sq): - """do a mql write""" - query = dict(query=sq, escape=False) - qstr = simplejson.dumps(query) - - self.log.debug('MQLWRITE: %s', qstr) - - service = '/api/service/mqlwrite' - - # should check log level to avoid redundant simplejson.dumps - self.log.info('%s: %s', - service, - simplejson.dumps(sq, indent=2)[1:-2]) - - r = self._httpreq_json(service, 'POST', - form=dict(query=qstr)) - - self.log.debug('MQLWRITE RESP: %r', r) - return self._mqlresult(r) - - def mqlflush(self): - """ask the service not to hand us old data""" - self.log.debug('MQLFLUSH') - - service = '/api/service/mqlwrite' - r = self._httpreq_json(service, 'POST', form={}) - - self._check_mqlerror(r) - return r - - def upload(self, body, content_type, document_id=False): - """upload to the metaweb""" - - service = '/api/service/upload' - - self.log.info('POST %s: %s (%d bytes)', - service, content_type, len(body)) - - - headers = {} - if content_type is not None: - headers['content-type'] = content_type - - form = None - if document_id is not False: - if document_id is None: - form = { 'document': '' } - else: - form = { 'document': document_id } - - # note the use of both body and form. - # form parameters get encoded into the URL in this case - r = self._httpreq_json(service, 'POST', - headers=headers, body=body, form=form) - return self._mqlresult(r) - - -if __name__ == '__main__': - console = logging.StreamHandler() - console.setLevel(logging.DEBUG) - - mss = HTTPMetawebSession('sandbox.freebase.com') - - self.mss.log.setLevel(logging.DEBUG) - self.mss.log.addHandler(console) - - - print mss.mqlread([dict(name=None, type='/type/type')]) diff --git a/freebase/rison.py b/freebase/rison.py deleted file mode 100755 index 2f4eebc..0000000 --- a/freebase/rison.py +++ /dev/null @@ -1,308 +0,0 @@ -# -# rison for python (parser only so far) -# see http://mjtemplate.org/examples/rison.html for more info -# - -###################################################################### -# -# the rison parser is based on javascript openlaszlo-json: -# Author: Oliver Steele -# Copyright: Copyright 2006 Oliver Steele. All rights reserved. -# Homepage: http:#osteele.com/sources/openlaszlo/json -# License: MIT License. -# Version: 1.0 -# - -# hacked by nix for use in uris -# ported to python by nix -# -# TODO -# -# switch to unicode -# fall through to simplejson if first char is not in '!(' - -# this allows code to use just one parser -# - - -import os, sys, re -#import simplejson -simplejson = None - -class ParserException(Exception): - pass - -class Parser(object): - WHITESPACE = '' - #WHITESPACE = " \t\n\r\f" - - # we divide the uri-safe glyphs into three sets - # and classes are illegal in ids. - # - used by rison (possibly later) - # - not common in strings, reserved - #not_idchar = "'!=:(),*@$;&"; - - idchar_punctuation = '_-./~' - not_idchar = ''.join([c for c in (chr(i) for i in range(127)) - if not (c.isalnum() - or c in idchar_punctuation)]) - - # additionally, we need to distinguish ids and numbers by first char - not_idstart = "-0123456789"; - - # regexp string matching a valid id - idrx = ('[^' + not_idstart + not_idchar + - '][^' + not_idchar + ']*') - - # regexp to check for valid rison ids - id_ok_re = re.compile('^' + idrx + '$', re.M) - - # regexp to find the end of an id when parsing - next_id_re = re.compile(idrx, re.M) - - def parse_json(self, str): - if len(str) > 0 and str[0] not in '!(': - return simplejson.loads(str) - return self.parse(str) - - def parse(self, str): - self.string = str - self.index = 0 - - value = self.readValue() - if self.next(): - raise ParserException("unable to parse rison string %r" % (str,)) - return value - - def readValue(self): - c = self.next() - - if c == '!': - return self.parse_bang() - if c == '(': - return self.parse_open_paren() - if c == "'": - return self.parse_single_quote() - if c in '-0123456789': - return self.parse_number() - - # fell through table, parse as an id - s = self.string - i = self.index-1 - - m = self.next_id_re.match(s, i) - if m: - id = m.group(0) - self.index = i + len(id) - return id # a string - - if c: - raise ParserException("invalid character: '" + c + "'") - raise ParserException("empty expression") - - def parse_array(self): - ar = [] - while 1: - c = self.next() - if c == ')': - return ar - - if c is None: - raise ParserException("unmatched '!('") - - if len(ar): - if c != ',': - raise ParserException("missing ','") - elif c == ',': - raise ParserException("extra ','") - else: - self.index -= 1 - n = self.readValue() - ar.append(n) - - return ar - - def parse_bang (self): - s = self.string - c = s[self.index] - self.index += 1 - if c is None: - raise ParserException('"!" at end of input') - if c not in self.bangs: - raise ParserException('unknown literal: "!' + c + '"') - x = self.bangs[c] - if callable(x): - return x(self) - - return x - - - def parse_open_paren (self): - count = 0 - o = {} - - while 1: - c = self.next() - if c == ')': - return o - if count: - if c != ',': - raise ParserException("missing ','") - elif c == ',': - raise ParserException("extra ','") - else: - self.index -= 1 - k = self.readValue() - - if self.next() != ':': - raise ParserException("missing ':'") - v = self.readValue() - - o[k] = v - count += 1 - - - def parse_single_quote(self): - s = self.string - i = self.index - start = i - segments = [] - - while 1: - if i >= len(s): - raise ParserException('unmatched "\'"') - - c = s[i] - i += 1 - if c == "'": - break - - if c == '!': - if start < i-1: - segments.append(s[start:i-1]) - c = s[i] - i += 1 - if c in "!'": - segments.append(c) - else: - raise ParserException('invalid string escape: "!'+c+'"') - - start = i - - - if start < i-1: - segments.append(s[start:i-1]) - self.index = i - return ''.join(segments) - - - # Also any number start (digit or '-') - def parse_number(self): - s = self.string - i = self.index - start = i-1 - state = 'int' - permittedSigns = '-' - transitions = { - 'int+.': 'frac', - 'int+e': 'exp', - 'frac+e': 'exp' - } - - while 1: - if i >= len(s): - i += 1 - break - - c = s[i] - i += 1 - - if '0' <= c and c <= '9': - continue - - if permittedSigns.find(c) >= 0: - permittedSigns = '' - continue - - state = transitions.get(state + '+' + c.lower(), None) - if state is None: - break - if state == 'exp': - permittedSigns = '-' - - self.index = i - 1 - s = s[start:self.index] - if s == '-': - raise ParserException("invalid number") - if re.search('[.e]', s): - return float(s) - return int(s) - - # return the next non-whitespace character, or undefined - def next(self): - l = len(self.string) - s = self.string - i = self.index - - while 1: - if i == len(s): - return None - c = s[i] - i += 1 - if c not in self.WHITESPACE: - break - - self.index = i - return c - - - bangs = { - 't': True, - 'f': False, - 'n': None, - '(': parse_array - } - - -def loads(s): - return Parser().parse(s) - -if __name__ == '__main__': - p = Parser() - - rison_examples = [ - "(a:0,b:1)", - "(a:0,b:foo,c:'23skidoo')", - "!t", - "!f", - "!n", - "''", - "0", - "1.5", - "-3", - "1e30", - "1e-30", - "G.", - "a", - "'0a'", - "'abc def'", - "()", - "(a:0)", - "(id:!n,type:/common/document)", - "!()", - "!(!t,!f,!n,'')", - "'-h'", - "a-z", - "'wow!!'", - "domain.com", - "'user@domain.com'", - "'US $10'", - "'can!'t'", - ]; - - for s in rison_examples: - print - print '*'*70 - print - print s - - print '%r' % (p.parse(s),) diff --git a/freebase/uritemplate.py b/freebase/uritemplate.py deleted file mode 100755 index 0a81147..0000000 --- a/freebase/uritemplate.py +++ /dev/null @@ -1,143 +0,0 @@ -# ======================================================================== -# Copyright (c) 2007, Metaweb Technologies, Inc. -# All rights reserved. -# -# Redistribution and use in source and binary forms, with or without -# modification, are permitted provided that the following conditions -# are met: -# * Redistributions of source code must retain the above copyright -# notice, this list of conditions and the following disclaimer. -# * Redistributions in binary form must reproduce the above -# copyright notice, this list of conditions and the following -# disclaimer in the documentation and/or other materials provided -# with the distribution. -# -# THIS SOFTWARE IS PROVIDED BY METAWEB TECHNOLOGIES AND CONTRIBUTORS -# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT -# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS -# FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL METAWEB -# TECHNOLOGIES OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, -# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, -# BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; -# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER -# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT -# LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN -# ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE -# POSSIBILITY OF SUCH DAMAGE. -# ======================================================================== - -# -# URI Templating in Python -# -# see http://bitworking.org/projects/URI-Templates/ -# and http://bitworking.org/news/URI_Templates -# -# note that this implementation may go away soon in -# favor of joe gregorio's own code: -# http://code.google.com/p/uri-templates/ -# -# -# this implementation can also parse URIs, as long as the -# template is sufficiently specific. to allow this to work -# the '/' character is forbidden in keys when parsing. -# later it should be possible to loosen this restriction. -# -# -# example: -# from whatever.uritemplate import expand_uri_template -# expand_uri_template('http://{host}/{file}', -# dict(host='example.org', -# file='fred')) -# -# TODO: -# allow parsing to be aware of http://www. and trailing / -# nail down quoting issues -# - -import os, sys, re -import urllib - -__all__ = ['expand_uri_template', 'URITemplate'] - - -def expand_uri_template(template, args): - """Expand a URI template using the given args dictionary. - """ - return URITemplate(template).run(args) - -def _uri_encode_var(v): - return urllib.quote(v, safe="-_.~!$&'()*+,;=:/?[]#@") - - -class URITemplate(object): - """a URITemplate is a URI with simple variable substitution. - """ - - VARREF = re.compile(r'\{([0-9a-zA-Z_]+)\}') - - def __init__(self, s): - """Compile a URITemplate from a string. - """ - self.template = s; - - self.params = [] - tsplit = self.VARREF.split(s) - rxs = ['^'] - for i in range(len(tsplit)): - if i % 2: - # track the vars used - self.params.append(tsplit[i]) - # vars match any string - # vars are assumed to lack '/' - this is imperfect... - rxs.append('([^/]*)') - else: - # quote special chars regexp interpretation - rxs.append(re.escape(tsplit[i])) - rxs.append('$') - self._parser = re.compile(''.join(rxs)) - - def __repr__(self): - return '' % self.template - - def run (self, args): - """Expand the template using the given args. - """ - def repl(m): - key = m.group(1) - return _uri_encode_var(args.get(key, '')) - uri = self.VARREF.sub(repl,self.template) - - #if self.parse(uri) is None: - # print 're-parsing generated uri failed: %r, %r' % (uri, self.template) - return uri - - - def parse(self, uri): - """Try to parse a URI, extracting variable values. - """ - m = self._parser.match(uri) - if m is None: - return None - return dict(zip(self.params, m.groups())) - - -if __name__ == '__main__': - # - # testcases are imported from the URI::Template module on CPAN - # - import urllib2, simplejson - fp = urllib2.urlopen('http://search.cpan.org/src/BRICAS/URI-Template-0.09/t/data/spec.json') - testsuite = simplejson.loads(fp.read()) - vars = dict([(k.encode('utf-8'),v) for k,v in testsuite['variables'].items()]) - nsucceed = 0 - nfail = 0 - for test in testsuite['tests']: - ut = URITemplate(test['template']) - uri = ut.run(vars) - if uri != test['expected']: - print 'FAILED %r expected %r' % (uri, test['expected']) - print ' vars: %r' % (vars,) - nfail += 1 - else: - nsucceed += 1 - print 'tests: %d succeeded, %d failed' % (nsucceed, nfail) -- cgit v1.3.1