Package to download historical data from Darwinex with different data format.
The dowload of data from Darwinex is performed via FTP, in each file there is information of each operation performed in the ticker. This package allows you to download the data in ohlc format ready for analysis with python with different periodicity.
- Requirements
- Installation
- Documentation and Tutorials
- Questions, Suggestions and Bugs
- Contributing / Development
- License
The following libraries are required before installing darwinexDownloader:
- Python 3.7+
- ftplib
- tempfile
- gzip
- pandas
- datetime
- os
The instructions to install this package are in the header of the (DarwinexDownloader pypi website)[https://test.pypi.org/project/darwinexDownloader/]
Once you have installed the library you have to import it in the source code:
import darwinexDownloader
This the library in your code, you can use darwinexDownloader to connect with Darwinex:
dwnx = darwinexDownloader.Connection('your_user', 'your_password') # clase
At this time, we are ready to download the data:
data = dwnx.download('EURUSD', "28-06-2022", "29-06-2022", '15Min')
Now in the 'data' variable we have the ohlc data ready to make our analysis.
| Date | open | high | low | close | volume |
|---------------------|---------|---------|---------|---------|-------------|
| 2022-06-28 00:00:00 | 1.05797 | 1.05850 | 1.05789 | 1.05810 | 952330000.0 |
| 2022-06-28 00:15:00 | 1.05812 | 1.05852 | 1.05807 | 1.05825 | 549950000.0 |
| 2022-06-28 00:30:00 | 1.05824 | 1.05853 | 1.05773 | 1.05779 | 503900000.0 |
| ... | ... | ... | ... | ... | ... |
| 2022-06-29 23:30:00 | 1.04417 | 1.04463 | 1.04411 | 1.04457 | 471505000.0 |
| 2022-06-29 23:45:00 | 1.04458 | 1.04463 | 1.04444 | 1.04456 | 239250000.0 |
In the download method we have 4 arguments.
The first one is the ticker. The valid tickers are show in the Darwinex website
The second and third arguments are dates. For the moment, date format must always be as follows: "dd-mm-yyyy" where the first date is the start date and the second is the end date.
And the 4th argument is the frecuency. This frecuency has the format used by pandas timeframes where the most used in trading can be:
- '1Min'
- '5Min'
- '15Min'
- '30Min'
- '1H'
- '4H'
- '1D'
- 'Week'
- 'Month'
Feel free to open an issue here.
Contributions welcome.
MIT License. Iván García-Ferreira