# TNRS

import sys

import time

import urllib2

# Config

initial_pause = 0.35 # sec

pause_growth_factor = 1.3

# Protocol params

#server = 'tnrs.iplantcollaborative.org' # live server

initial_headers = {

    'Content-Type': 'text/x-gwt-rpc; charset=utf-8',

    'X-GWT-Module-Base': url_base,

    'X-GWT-Permutation': '574AA16D15D917C7704646FD92AFF6B3',

}

submission_request_template = ('7|0|7|'+url_base+

"classification":"tropicos", "match_to_rank":"true"}|0.05|1|2|3|4|2|5|5|6|7|')

retrieval_request_template = ('7|0|15|'+url_base+

'|1E87C78041CEFBF0992F46BDF84D7D60|org.iplantc.tnrs.demo.client.SearchService\

|getRemoteData|com.extjs.gxt.ui.client.data.PagingLoadConfig|\

java.lang.String/2004016611|com.extjs.gxt.ui.client.data.BasePagingLoadConfig/\

2011366567|com.extjs.gxt.ui.client.data.RpcMap/3441186752|sortField|sortDir|\

com.extjs.gxt.ui.client.Style$SortDir/640452531|offset|java.lang.Integer/\

3438268394|limit|{"email":"tnrs@lka5jjs.orv", "key":"[key]", \

	#  only from within the matched higher ranks. must be turned on, to ensure

	#  that higher ranks are always authoritative.

	#  the sources, regardless of match score. should never be turned on,

	#  because it selects worse matches instead of better ones. *however*, since

	#  this is currently broken and always forced on, we turn it on so that the

	#  download settings reflect what TNRS actually used.

BasePagingLoadResult/496878394","java.util.ArrayList/4159755760","org.iplantc.\

tnrs.demo.shared.BeanTNRSEntry/1039545748",".*"\],0,7\]$'

retrieval_response_info_pattern = r'(?ms).*^Set-Cookie: JSESSIONID=(\w+);'

download_request_template = ('7|0|6|'+url_base+

'|1E87C78041CEFBF0992F46BDF84D7D60|org.iplantc.tnrs.demo.client.SearchService|\

downloadRemoteResults|java.lang.String/2004016611|{"name":"tnrs_results.txt", \

	#  when on, sometimes marks multiple names as Selected. must be turned *off*

	#  to be on in the web app (see r9910, 2013-6-18), but does not appear to be

	#  needed (the source_sorting bug alluded to in r9910 is not fixed by

	#  enabling the dirty setting).

download_url_suffix = '&name=tnrs_results.txt&encoding=utf8'

	'Name_matched_author': 'Canonical_author',

	'Accepted_name_species': 'Accepted_species',

	'Accepted_name_family': 'Accepted_family',

	'Selected': 'selected',

}

    # embedded strs within GWT sections additionally have " JSON-escaped

    return strings.json_decode(strings.unesc_quotes(str_, '|', quote_esc='\!'))

    return [(r'(?: |(?<=\t)|^)'+re.escape(from_.strip())+r'(?: |(?=\t)|$)',

        strings.regexp_repl_esc(to)) for from_, to in decode_map]

    ('\t', ' !tab '), # TNRS replaces with " "

    ('\n', ' !nl '), # used to separate multiple names

    ('\r', ' !cr '), # used to separate multiple names

    ('\\', ' !bsl '), # TNRS removes it

decode_map = strings.flip_map(encode_map)

    str_ = strings.replace_all(encode_map, str_)

    # Empty and whitespace-only strings are ignored by TNRS (no response row)

    if str_.strip() == '': str_ = padding+str_

    return str_

def decode(str_): return strings.replace_all_re(spliced_decode_map, str_)

decode_for_tsv_map = make_spliced_decode_map([(from_, strings.replace_all(

    csvs.tsv_encode_map, to)) for from_, to in decode_map])

def decode_for_tsv(str_):

    return strings.replace_all_re(decode_for_tsv_map, str_)

class TnrsOutputStream(streams.FilterStream):

    '''Decodes a TNRS response whose names were encoded with encode()'''

    def __init__(self, stream):

        streams.FilterStream.__init__(self, decode_for_tsv, stream)

    match = re.match(pattern, str_)

    if not match:

        raise InvalidResponse('Invalid '+name+' response:\n'+response_info+'\n'

            +response)

    Note that names containing only whitespace characters (after gwt_encode())

    are ignored by TNRS and do not receive a response row. Thus, you should

    always match up the Name_submitted returned by TNRS with the actual

    submitted name to determine the corresponding TNRS response row.

    @param use_tnrs_export whether to use TNRS's TSV export feature, which

        currently returns incorrect selected matches (vegpath.org/issues/943)

    assert name_ct <= max_names

    # Logging

    def debug_log(label, str_=''):

        if debug: sys.stderr.write('\n'+label+':\n'+str_+'\n')

        debug_log('fetching export columns')

        export_cols = csvs.header(single_tnrs_request([''], debug,

            use_tnrs_export=True)) # need at least 1 name

        json_cols = [tsv_to_json_col_map.get(c, c) for c in export_cols]

        debug_log('JSON columns', strings.urepr(json_cols))

        debug_log('request str', str(request_obj.get_data()))

        response = urllib2.urlopen(request_obj)

        response_info = str(response.info())

        debug_log('response info', response_info)

        debug_log('response str', response_str)

        return response_str, response_info

    def do_repeated_request(request):

        pause = initial_pause

        total_pause = 0

        while True:

            total_pause += pause

            if total_pause > max_pause: raise # error is not temporary

            debug_log('total_pause', str(total_pause)+'s')

            time.sleep(pause) # wait for job to complete

            try: return do_request(request)

            except urllib2.HTTPError: pass # try again

            pause *= pause_growth_factor

    try:

        debug_log('Submit')

            response, response, response_info)

        key_enc = gwt_encode(key)

        debug_log('Retrieve')

        request = retrieval_request_template.replace('[key]', key_enc)

        response, response_info = do_repeated_request(request)

            response, response_info)

        session_id, = parse_response('retrieval info',

            retrieval_response_info_pattern, response_info, response,

        debug_log('session_id', session_id)

        headers['Cookie'] = 'JSESSIONID='+session_id

        # contain all the columns, contains no column names, and has different

        # lengths depending on the taxonomic ranks present in the provided taxon

        # name. the extra download step is therefore necessary.

            debug_log('Prepare download')

            request = download_request_template.replace('[key]', key_enc)

            response, response_info = do_request(request)

            csv_url, = parse_response('download', download_response_pattern,

                response, response, response_info)

            csv_url += download_url_suffix

            debug_log('csv_url', csv_url)

            debug_log('Download')

            request_obj = urllib2.Request(csv_url)

            debug_log('request URL', str(csv_url))

            debug_log('request info', str(request_obj.header_items()))

            debug_log('request str', str(request_obj.get_data()))

            response = urllib2.urlopen(request_obj)

            response_info = str(response.info())

            debug_log('response info', response_info)

            debug_log('Expand')

            request = ('7|0|6|http://tnrs.iplantcollaborative.org/tnrsdemo/|\

1E87C78041CEFBF0992F46BDF84D7D60|org.iplantc.tnrs.demo.client.SearchService|\

requestGroupMembers|java.lang.String/2004016611|{"group":"0", \

"email":"tnrs@lka5jjs.orv", "key":"[key]", "source_sorting":"false", \

"taxonomic_constraint":"false"}|1|2|3|4|1|5|6|'.replace('[key]', key_enc))

            response, response_info = do_request(request)

            data_gwt, = parse_response('expand', r'^//OK\[1,\["(.*)"\],0,7\]$',

                response, response, response_info)

            data_json = json.loads(gwt_decode(data_gwt))

            output_stream = csvs.InputRewriter(

                csvs.RowNumFilter(

                )

            )

        if cumulative_profiler != None:

            cumulative_profiler.add_subprofiler(profiler)

            sys.stderr.write('Cumulatively: '+cumulative_profiler.msg()+'\n')

            # try again with debug turned on