20 1 1 6

TRTHv2 speedguide download is slow

I tried to use the REST API to retrieve Speedguide pages, as explained in the user guide (v3.0) of Thomson Reuters Tick History 11.1 REST API Chapter 5. After figuring out that the error in the first reply can be ignored and I can go ahead looking for the ExtractionID and extract files in it, I can successfully get the data of the page as described in the manual. However, I found that it takes a really long time to get the page. When fetching manually, I usually get the page in 10 seconds. But when fetched with the REST API, it usually takes between 30 seconds to 1 minute in order to get a Speedguide page, in particular for the Extractions/ExtractRaw call to complete. Is there anything I can do to fetch the data faster? I attach my complete Python test program for reference.

import datetime
import getpass
import json
import os
import re
import time

import requests

def make_dss_odata(od_type, **kwargs):
    "Make a dict object as DSS odata"
    kwargs.update({'@odata.type': '#ThomsonReuters.Dss.Api.' + od_type})
    return kwargs

class ReutersDSSCaller(object):
    "Make REST API call to Reuters DataScope Select"
    _BASEURL = ''

    def __init__(self, user, pw):
        self._user = user
        self._pw = pw
        self._auth_token = None

    def get(self, *args, **kwargs):
        "Call REST get method"
        return self._call('get', *args, **kwargs)

    def put(self, *args, **kwargs):
        "Call REST put method"
        return self._call('put', *args, **kwargs)

    def delete(self, *args, **kwargs):
        "Call REST delete method"
        return self._call('delete', *args, **kwargs)

    def post(self, *args, **kwargs):
        "Call REST post method"
        return self._call('post', *args, **kwargs)

    def _call(self, method, path, headers=(), body=None, parse=3, stream=False):
        # pylint: disable=R0913
        headers = dict(headers)
        headers.setdefault('Authorization', 'Token ' + self.get_auth_token())
        return self._raw_call(method, path, headers, body, parse, stream)

    def get_auth_token(self):
        "Get DSS REST API authentication token"
        if not self._auth_token:
            ret = self._raw_call('post', 'Authentication/RequestToken', {}, {
                'Credentials': {
                    'Password': self._pw,
                    'Username': self._user
            }, 3, False)
            self._auth_token = ret
        return self._auth_token

    def _raw_call(self, method, path, extraheaders, body, parse, stream):
        # pylint: disable=R0913
        headers = {'Content-Type': 'application/json; odata.metadata=minimal'}
        body_str = '' if body is None else json.dumps(body)
        resp = getattr(requests, method)(self._BASEURL + '/' + path,
                                         data=body_str, headers=headers,
        if parse <= 0:
            return resp
        if resp.status_code >= 400:
            raise RuntimeError('DSS API Error %s: %s'
                               % (resp.status_code, resp.reason))
        if parse <= 1 or stream:
            return resp
        ret = resp.json()
        return ret if parse <= 2 else ret['value']

    def download(self, resp, path):
        "Download content in REST API response"
        tmp_path = path + '.tmp'
        with open(tmp_path, 'wb') as fout:
            if path.endswith('.gz'):
                content_iter = self._iter_raw_resp(resp, 1024 * 1024)
                content_iter = resp.iter_content(chunk_size=1024 * 1024)
            for chunk in content_iter:
        os.rename(tmp_path, path)

    def _iter_raw_resp(self, resp, chunk_size):
        while True:
            chunk =
            if len(chunk) == 0:
            yield chunk

# Create caller
user_name = raw_input('DSS User name: ')
passwd = getpass.getpass('DSS password: ')
caller = ReutersDSSCaller(user_name, passwd)
page = raw_input('Speedguide page: ')
if not page:
    page = 'CBT/TY'
    print 'Using default,', page

# Submit extraction request
today = datetime.datetime.combine(, datetime.time())
start_date = today - datetime.timedelta(7)
print 'Initial call'
ret =
    headers={'Prefer': 'respond-async'},
        'ExtractionRequest': make_dss_odata(
                ValidationOptions={'AllowHistoricalInstruments': True},
                    {'Identifier': page,
                     'IdentifierType': 'Ric'}
                'MessageTimeStampIn': 'GmtUtc',
                'ReportDateRangeType': 'Range',
                'QueryStartDate': start_date.isoformat(),
                'QueryEndDate': today.isoformat(),
                'ExtractBy': 'Ric',
                'SortBy': 'SingleByRic',
                'DomainCode': 'MarketPrice',
                'DisplaySourceRIC': True

# Poll for completion
if ret.status_code == 202:
    loc = ret.headers['location'].partition('/v1/')[2]
    while ret.status_code == 202:
        print 'Initial call retry'
        ret = caller.get(loc, headers={'Prefer': 'respond-async'}, parse=1)
print 'Initial call completed'

# Look for extraction ID
match ='Extraction ID: ([0-9]+)\n', ret.json()['Notes'][0])
eid =

# List Report Files
file_list = caller.get('Extractions/ReportExtractions(\'%s\')/Files' % eid)

# Download the Files
for f_spec in file_list:
    filename = f_spec['ExtractedFileName']
    ret = caller.get('Extractions/ExtractedFiles(\'%s\')/$value'
                     % f_spec['ExtractedFileId'], stream=True)
    print 'Downloading', filename, filename)

icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

56.3k 145 45 65

From the note file (_OnD_0xxxx.csv.gz.notes.txt), you can see how long the server processes the extraction.

Processing started at 07/26/2017 13:56:07
Processing completed successfully at 07/26/2017 13:56:33
Extraction finished at 07/26/2017 06:56:33 UTC, with servers: tm07n01

However, getting the completed status via REST API can be longer than that. I have added the time stamp in the code before sending the request and after receiving the response.

('Request Timestamp: 2017-07-26 13:56:07', 'Extractions/ExtractRaw')
('Response Timestamp: 2017-07-26 13:56:37', 'Extractions/ExtractRaw')
Initial call completed

The server used 26 seconds to process the extraction. For the python application, it used 30 seconds to get the completed status.

The result is similar to Postman. The following is the output of request from Postman.

From the result, the server used 28 seconds to process the extraction and it used around 32 seconds for Postman to get the response back.

Typically, it is possible to take more than 30 seconds to process the extraction. It depends on the server load at that time.

postman.png (57.6 KiB)
icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

This is from a note file of an immediate schedule via GUI.

Processing started at 07/26/2017 16:21:11
Processing completed successfully at 07/26/2017 16:21:38

It took 27 seconds to complete the process.

20 1 1 6

So it means there is no way to get the experience of the web UI, where a speedguide page can be fetched in matter of a few seconds?

icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

Is it View Page link on GUI?

If yes, I couldn't find this similar feature in REST API.

Moreover, it is mentioned in REST API guide that to retrieve a Speed Guide page via the Tick History REST API, you need to use the ExtractRaw endpoint against the Raw report template.

20 1 1 6

You are right, by "the experience of the web UI" I'm referring to the ability to fetch the page quickly by searching for the speed guide name using Historical search, clicking the "view page" button and following the links within the pages.

The Extractions/ExtractRaw endpoint is used in my program, with the ExtractionRequest parameter set as "TickHistoryRawExtractionRequest". But I don't understand what you mean by using the endpoint against the Raw report template. The example shown in the user guide doesn't seem to be specifying a report template. The API reference tree doesn't seem to indicate a template parameter either.

icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.

TickHistoryRawExtractionRequest is a Raw report template mentioned in the guide.

20 1 1 6

I see. So I'm already doing the best I can, and can only hope that Reuters will provide a better method to get speedguide pages in the future.

icon clock
10 |1500

Up to 2 attachments (including images) can be used with a maximum of 512.0 KiB each and 1.0 MiB total.