mirror of
https://github.com/OpenBB-finance/OpenBB.git
synced 2026-10-27 12:40:35 +00:00
58
README.md
58
README.md
@@ -132,6 +132,21 @@ This is a library for package management, and ensures a smoother experience than
|
||||
python terminal.py
|
||||
```
|
||||
|
||||
### Advanced User Install - Python 3.8
|
||||
|
||||
*Note that the `conda deactivate` -> `conda activate` in the middle is on purpose, this is sometimes required to avoid issues with poetry*
|
||||
|
||||
```
|
||||
conda create -n gst python=3.8.8
|
||||
conda activate gst
|
||||
conda install poetry
|
||||
conda deactivate
|
||||
conda activate gst
|
||||
poetry install
|
||||
conda install -c conda-forge fbprophet numpy=1.19.5 hdf5=1.10.5
|
||||
poetry install -E prediction
|
||||
```
|
||||
|
||||
### Advanced User Install - Machine Learning
|
||||
|
||||
If you are an advanced user and use other Python distributions, we have several requirements.txt documents that you can pick from to download project dependencies.
|
||||
@@ -145,10 +160,17 @@ Note: The libraries specified in the [requirements.txt](/requirements.txt) file
|
||||
ENABLE_PREDICT = os.getenv("GTFF_ENABLE_PREDICT") or True
|
||||
```
|
||||
|
||||
* Install optional ML features dependencies collection with poetry:
|
||||
* Install optional ML features dependencies:
|
||||
```
|
||||
poetry install -E prediction
|
||||
```
|
||||
*If you run into issues installing or `Cannot convert a symbolic Tensor...` at runtime, try this:*
|
||||
|
||||
```
|
||||
conda install -c conda-forge fbprophet numpy=1.19.5 hdf5=1.10.5
|
||||
poetry install
|
||||
poetry install -E prediction
|
||||
```
|
||||
|
||||
*If you would like to set up a docker image:*
|
||||
|
||||
@@ -157,10 +179,40 @@ poetry install -E prediction
|
||||
|
||||
Note: The problem with docker is that it won't output matplotlib figures.
|
||||
|
||||
*Commands that may help you in case of an error:*
|
||||
* `conda install -c conda-forge fbprophet -y`
|
||||
*Commands that may help you in case of an error:
|
||||
|
||||
* `python -m pip install --upgrade pip`
|
||||
* `pip install pystan --upgrade`
|
||||
* `poetry update --lock`
|
||||
|
||||
### Other issues
|
||||
|
||||
If you run into trouble with poetry and the advice above did not help, your best bet is to try
|
||||
|
||||
1. `poetry update --lock`
|
||||
|
||||
2. `conda deactivate` -> `conda activate gst`, then try again
|
||||
|
||||
3. Delete the poetry cache, then try again
|
||||
|
||||
| Platform | Location |
|
||||
| -------- | ------------------------------- |
|
||||
| Linux | "~/.cache/pypoetry" |
|
||||
| Mac | "~/Library/Caches/pypoetry" |
|
||||
| Windows | "%localappdata%/pypoetry/cache" |
|
||||
|
||||
4. Track down the offensive package and purge it from your anaconda `<environment_name>` folder, then try again (removing through conda can sometimes leave locks behind)
|
||||
|
||||
| Platform | Location |
|
||||
| --------- | -------------------------------------------- |
|
||||
| Linux/Mac | "~/anaconda3/envs" or "~/opt/anaconda3/envs" |
|
||||
| Windows | "%userprofile%/anaconda3/envs" |
|
||||
|
||||
5. Completely nuke your conda environment folder and make a new environment from scratch
|
||||
|
||||
6. Reboot your computer and try again
|
||||
|
||||
7. Submit a ticket on github
|
||||
|
||||
### API Keys
|
||||
|
||||
|
||||
Reference in New Issue
Block a user