在Conda环境中安装Apache-Airflow

Question

The Situation

I am trying to install apache-airflow using pip into a conda environment. I am not able to, since setup.py egg_info keeps failing:

Command "python setup.py egg_info" failed with error code 1 in C:\\Users\\joshu\\AppData\\Local\\Temp\\pip-install-3efyslfh\\apache-airflow\\

message when I run

pip install "apache-airflow[s3, postgres]"

I ran this in Windows cmd, then did all the steps again as admin in cmd (shift+enter when opening cmd).

(myVenv) C:\Users\joshu\Documents>pip install "apache-airflow[s3, postgres]"
Collecting apache-airflow[postgres,s3]
  Using cached https://files.pythonhosted.org/packages/9e/12/6c70f9ef852b3061a3a6c9af03bd9dcdcaecb7d75c8898f82e3a54ad5f87/apache-airflow-1.9.0.tar.gz
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "C:\Users\joshu\AppData\Local\Temp\pip-install-3efyslfh\apache-airflow\setup.py", line 102
        async = [
              ^
    SyntaxError: invalid syntax

    ----------------------------------------
Command "python setup.py egg_info" failed with error code 1 in C:\Users\joshu\AppData\Local\Temp\pip-install-3efyslfh\apache-airflow\

Steps I took

1) I first created a conda environment and installed pip and setuptools into the environment:

C:\Users\joshu\Documents>conda create -n airflow pip setuptools
Solving environment: done


==> WARNING: A newer version of conda exists. <==
  current version: 4.5.4
  latest version: 4.5.10

Please update conda by running

    $ conda update -n base conda



## Package Plan ##

  environment location: C:\Users\joshu\Anaconda3\envs\airflow

  added / updated specs:
    - pip
    - setuptools


The following NEW packages will be INSTALLED:

    certifi:        2018.8.13-py37_0
    pip:            10.0.1-py37_0
    python:         3.7.0-hea74fb7_0
    setuptools:     40.0.0-py37_0
    vc:             14.1-h0510ff6_3
    vs2015_runtime: 15.5.2-3
    wheel:          0.31.1-py37_0
    wincertstore:   0.2-py37_0

C:\Users\joshu\Documents>activate airflow

(airflow) C:\Users\joshu\Documents> pip freeze
certifi==2018.8.13
wincertstore==0.2

(airflow) C:\Users\joshu\Documents>conda list
# packages in environment at C:\Users\joshu\Anaconda3\envs\airflow:
#
# Name                    Version                   Build  Channel
certifi                   2018.8.13                py37_0
pip                       10.0.1                   py37_0
python                    3.7.0                hea74fb7_0
setuptools                40.0.0                   py37_0
vc                        14.1                 h0510ff6_3
vs2015_runtime            15.5.2                        3
wheel                     0.31.1                   py37_0
wincertstore              0.2                      py37_0

(airflow) C:\Users\joshu\Documents>where pip
C:\Users\joshu\Anaconda3\envs\airflow\Scripts\pip.exe
C:\Users\joshu\Anaconda3\Scripts\pip.exe

(airflow) C:\Users\joshu\Documents>which pip
/c/Users/joshu/Anaconda3/envs/airflow/Scripts/pip

After this setup, running pip install apache-airflow gives me the error at the top of this page

Other steps I took

1. Installing apache-airflow on my system (no conda environment). This works, but I'm looking to have it contained in an environment.
2. Upgrading setuptools:

code:

(airflow) C:\Users\joshu\Documents>python -m pip install --upgrade pip
Collecting pip
  Using cached https://files.pythonhosted.org/packages/5f/25/e52d3f31441505a5f3af41213346e5b6c221c9e086a166f3703d2ddaf940/pip-18.0-py2.py3-none-any.whl
Installing collected packages: pip
  Found existing installation: pip 10.0.1
    Uninstalling pip-10.0.1:
      Successfully uninstalled pip-10.0.1
Successfully installed pip-18.0

(airflow) C:\Users\joshu\Documents>which pip
/c/Users/joshu/Anaconda3/envs/airflow/Scripts/pip

(airflow) C:\Users\joshu\Documents>pip --version
pip 18.0 from c:\users\joshu\anaconda3\envs\airflow\lib\site-packages\pip (python 3.7)

(airflow) C:\Users\joshu\Documents>pip install "apache-airflow[s3, postgres]"

I've also tried upgrading pip then running install -- no luck

 (airflow) C:\\Users\\joshu\\Documents>python -m pip install --upgrade pip Collecting pip Using cached https://files.pythonhosted.org/packages/5f/25/e52d3f31441505a5f3af41213346e5b6c221c9e086a166f3703d2ddaf940/pip-18.0-py2.py3-none-any.whl Installing collected packages: pip Found existing installation: pip 10.0.1 Uninstalling pip-10.0.1: Successfully uninstalled pip-10.0.1 Successfully installed pip-18.0 (airflow) C:\\Users\\joshu\\Documents>which pip /c/Users/joshu/Anaconda3/envs/airflow/Scripts/pip (airflow) C:\\Users\\joshu\\Documents>pip --version pip 18.0 from c:\\users\\joshu\\anaconda3\\envs\\airflow\\lib\\site-packages\\pip (python 3.7) (airflow) C:\\Users\\joshu\\Documents>pip install "apache-airflow[s3, postgres]" 

Ultimately I would like to have airflow and related packages I need for my project in a conda environment.

Answer 1

The error appears to be because async is a reserved keyword in Python 3.7 and this package has not updated their setup.py script. You can use Python 3.6 until the package is updated

conda create -n airflow pip setuptools python=3.6
conda activate airflow
pip install "apache-airflow[s3, postgres]"

Answer 2

I also had the same problem. I solved it by installing with conda instead of pip:

conda install -c conda-forge airflow

https://anaconda.org/conda-forge/airflow

Answer 3

I ran into the same problems. Apache Airflow 1.10 is incompatible.

My solution was to build the source code. I got the latest source code via git. This command creates a wheel file in the dist directory.

python3 setup.py bdist_wheel

Then you install

pip3 install apache_airflow-2.0.0.dev0+incubating-py3-none-any.whl