This package seems to no longer work; I'm getting these error messages about duka

[milo-hyben]

Try to limit number of threads default is 10, I am using 2 threads and it seems to work, anything above 2 is throwing the Exception as on the first comment.
Example of working call:

"duka EURUSD -s 2017-01-01 -e 2017-12-31 -t 2"

[smorrel1]

+1 on this. Any help appreciated please :)

I wonder if Dukascopy started requesting a logon recently which broke this? The behaviour when going incognito supports this theory, i.e. when pressing 'Download' incognito mode for the first time you're asked to log on, but not on subsequent downloads. See this note on incognito.
res = await loop.run_in_executor(None, lambda: requests.get(url, stream=True)) in fetch.py line 24 doesn't send cookies. Here are the docs, but I've not found a way to send them yet.

However sometimes this code has worked for 1 day at a time, so I'm not sure....

[cjlv91]

Getting the same errors, anyone have potential solutions?

Anyone with sophisticated enough knowledge maybe can incorporate logins into their system? I.e. maybe bypassing the errors using accountlogin. I just registered an account (not trading account) solely to download the data, and it works wonderfully downloading directly from their website.

[smorrel1]

What about logging in manually then including the session cookies with the requests.get()?

[cjlv91]

Well it seems that the problem might lie at Dukascopy's own servers, I cannot download data for more than 1 day back on their website using proper login manually. For instance I try download Brent 15min data for 4months. Doesn't load, just stuck at "50%" loading for hours now. This seem to occur at the same time the python give those errors of no data fetched.

Might be something with their servers?

Is there a limit on how much data you can download from their website using proper login manually?

[smorrel1]

I could download data for 2003 in some pairs, and two weeks at a time from a month back. But I got asked to log in at apparently random intervals. Which means that the server is fine, but logging in is the problem, provided the downloads are no longer than say 10 days each. Possibly they introduced throttling as well as logon to thwart bulk downloads.

If anyone knows how to get cookies from Chrome’s SQL database then I can amend the code to send these and do a pull request.

[nick2012]

@smorrel1 can you please implement it for firefox cookies? just add an argument for the cookiefile path. I've tried to do it myself but I'm not that advanced, and I could not make it to work. But I guess this is the way to do it: https://stackoverflow.com/questions/49502254/how-to-import-firefox-cookies-to-python-requests

[smorrel1]

Hi, after some investigation I might have missed diagnose the problem... you could try using JFOREX to download historic data quite easily instead, it’s what worked best for me.

[nick2012]

@smorrel1 Can you please guide me on how to use JFOREX? I need stock data feed.

[tijo45]

@milo-hyben thanks! that worked for me.

[silviojaeger]

@nick2012 JFOREX workes fine to download historical data. Just open a demo account on dukascopy, download and install JFOREX and sign in with your demo account.
In the menubar click on "View" --> "Historical Data Manager".
On the popup window you can make your settings and download everything, it's pretty straight forward.
I downloaded the complet Dukascopy dataset in a few hours.

[hn2]

Does it still work for anyone? I get Request failed ... even with 1 thread

[alanpang1990301]

not working

[ipray4]

Try to limit number of threads default is 10, I am using 2 threads and it seems to work, anything above 2 is throwing the Exception as on the first comment.
Example of working call:

"duka EURUSD -s 2017-01-01 -e 2017-12-31 -t 2"

wow it really works... Thanks a lot