The ice2oceans_api from whigg

<!DOCTYPE html>

<html lang="en" xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta charset="utf-8" />
<title>Ice2Ocean Server setup and install readme</title>
</head>
<body>
<pre>
First things first, you must install the current python version:

-- Python 2.7.8 (default, Jun 30 2014, 16:03:49) [MSC v.1500 32 bit (Intel)] on win32
-- Plus some python modules but these will be installed in the "virtual environment" step below

Second:

-- You will need a virtual environment.
-- You must construct your own as 'pre-built' is not an option at this juncture (9/2014).

To construct your own virtual environment:

Prereq:
You will need this version of python (go to link and click install): https://www.python.org/ftp/python/2.7.8/python-2.7.8.msi.
This version of the cpython interpreter works reliably with Azure Websites as of (9/2014).
It also works reliably with cloud services and VM's.
As you install: Be very "YES DO IT" in your approach but WARNING: the Customize Python 2.7.8 form on the wizard has a red X at the
option to "Add python.exe to Path". Select the pull-down choice "will be installed on local hard drive". Proceed.
You will need the latest version of Python Tools for Visual Studio. This installation guide assumes you can do that!

A moment of explanation: As we add References to C# projects we add (similar idea) a python runtime environment to VS python Projects as needed.
In particular we have the LiveOceanServer Project (notice icon has "PY" in the box) and this will need a virtual environment inside of which
will exist the python runtime. So let's do that (an easy way of getting the python stuff running in Azure).

In the VS python project 'LiveOceanServer' right click on the Python Environments section and select "Add Virtual Enviroment".
Follow along with the wizard: env, python 2.7, box checked. This installs some 'easy to do' dependencies

Right click on the environment you have created "env (Python 2.7)". Select 'Open Command Prompt Here...'.

At the prompt type:

> Scripts\activate

Another moment of explanation: Notice in your command console that the prompt now starts with (env) if you are in good shape. From now on
when you install packages (python utilities / libraries like NumPy or SciPy) they will be installed within this virtual environment, NOT
in your global python environment. So this is very much a sandbox.

Now fetch each binary installer individually from: http://www.lfd.uci.edu/~gohlke/pythonlibs/,
with the notable exception of pyproj. Here we have some specific instructions:
- Cut and paste this URL into a browser: https://code.google.com/p/pyproj/downloads/detail?name=pyproj-1.9.3.win32-py2.7.exe&can=2&q=
- Save the install file (do not run it) to a folder like Downloads
- Go to that folder, right click on the install file: Properties
- Click "Unblock" so it is greyed out, then Apply, Ok
- Leave Windows Explorer, return to Visual Studio
- Locate the LiveOceanServer Project (not the entire Solution) and open Python Environments; notice there is "env (Python 2.7)" here
- Right click on "env (Python 2.7)" and select Open Command Prompt Here
- In the Command Prompt window type in Scripts\activate.bat and return
- Notice that the prompt returns starting with (env); this means that the virtual environment (env) has been activated
- This means pip and easy_install will attempt to install things into the virtual python environment rather than the global python environment
- Change directories in the Command Prompt window to where the python installer sits
- easy_install pyproj-1.9.3-py2.7-win32.exe
- If this finishes then it will appear in the listing for the environment in VS
- If this does not work please see further documentation just below.

Packages Required:
* python-dateutil better datetime support than default
* six a dependency; not needed directly (gridded data)
* tornado a dependency; not needed directly (gridded data)
* pandas a dependency; not needed directly (gridded data)
* Numpy matrix operations against gridded data (and more)
* Scipy a dependency; not needed directly
* NetCDF4 provides access to content of NetCDF files
* matplotlib makes pngs from data (and many other things)
* pyproj bing maps projection

There will be options for several versions of python. Make sure you get the one specifically for python 2.7 and win32.
This is because we are using this on Azure in a machine that is 32bit. We could use a 64 bit machine in Azure; the important thing
is that the python version matches what-bits-we-use in Azure.

Now move each of the binary (*.exe) files you've just grabbed into the virtual enviroment folder.

Back at the command prompt in you virtual enviroment (the one we set up at the start; if its been closed, repeat the bit about activating the virtual enviroment):
Anthony 20141008: or run the "installythonPackages.bat" file that has all of this in one place

(env) ... > easy_install python-dateutil-2.2.win32-py2.7.exe
(env) ... > easy_install six-1.7.3.win32-py2.7.exe
(env) ... > easy_install tornado-4.0.1.win32-py2.7.exe
(env) ... > easy_install numpy-MKL-1.8.2.win32-py2.7.exe
(env) ... > easy_install pandas-0.14.1.win32-py2.7.exe
(env) ... > easy_install scipy-0.14.0.win32-py2.7.exe
(env) ... > easy_install netCDF4-1.1.1.win32-py2.7.exe
(env) ... > easy_install matplotlib-1.4.0.win32-py2.7.exe
(env) ... > easy_install pyproj-1.9.3.win32-py2.7.exe
(env) ... > easy_install psycopg2-2.5.4.win32-py2.7.exe

Anthony 20141008: I found these did not install as part of the "easy to do" dependancies described above. I included them in the batch file

(env) ... > easy_install azure
(env) ... > easy_install django==1.6.5
(env) ... > easy_install django-filter
(env) ... > easy_install docutils
(env) ... > easy_install markdown

And your virtual enviroment should now be set up for running LiveOceanServer. To check: in Visual Studio hit "Play" (green triangle where it says
Internet Explorer: It is going to use IE to show you the server on your localhost and a default Client so you can see what it is doing.)

Anecdotal: Rob on 9/5/2014 run-through had to install Scipy a second time; first time through the launch failed.

Pyproj installation "version 2" for when version 1 (above) does not work
Potential Pitfall: Pyproj has in various instances failed to properly install -- even with easy_install and an administrative console.
If this happens, don't yell at the heavens and take out your frustration on helpless inanimate objects.

One tack that seems to work reliably is to install it in the global site packages folder.

Do not invoke Scripts\activate in the console,
Navigate to the location of the pyproj-1.9.3.win32-py2.7.exe executable

... > easy_install pyproj-1.9.3.win32-py2.7.exe

Now, in your python 2.7 installation directory locate the folder: Libs\site-packages\pyproj-1.9.3-py2.7-win32.egg
Copy the entire directory.
In VS right click the virtual environment you intend to use (env) and select Open Folder in File Explorer.
Navigate to from the root directory of the virtual enviroment to Lib\site-packages and paste the copied directory.
Finally, in the same directory locate a file called easy_install.pth
Open this file in the text editor of your choice.

You will see something like this:

import sys; sys.__plen = len(sys.path)
./numpy-1.8.2-py2.7-win32.egg
./netcdf4-1.1.0-py2.7-win32.egg
./pandas-0.14.1-py2.7-win32.egg
./scipy-0.14.0-py2.7-win32.egg
./matplotlib-1.3.1-py2.7-win32.egg
./pyparsing-2.0.2-py2.7-win32.egg
./tornado-4.0.1-py2.7-win32.egg
./backports.ssl_match_hostname-3.4.0.2-py2.7.egg
./certifi-14.05.14-py2.7.egg
import sys; new=sys.path[sys.__plen:]; del sys.path[sys.__plen:]; p=getattr(sys,'__egginsert',0); sys.path[p:p]=new; sys.__egginsert = p+len(new)

Add './pyproj-1.9.3-py2.7-win32.egg' on a new line before the final 'import ...'
Save this file and restart VS.

Pyproj should now appear in the list of installed packages in the virtual enviroment.

Step 3: See if it works
In VS set LiveOceanServer as the default project
The >Start button now reads >Internet Explorer. Click this.
IE should start with http://localhost:XXXXX (changes each time you run)
You should see the LiveOcean splash screen
Add /admin to the URL and log in to Django administration using admin and the password, whatever that is

Step 4: Creating a new LiveOceanServer super user
In VS look at the LiveOceanServer Solution: Here you will find the LiveOceanServer Project.
Within this Project is (unless you are starting fresh) a file called db.sqlite3.
This file contains super user information and other Django configuration info
You can delete it and regenerate it as needed; but you do need it to exist to do Django administration.
Stop the instance running.
Open the LiveOceanServer Project folder in Windows Explorer and BE SURE THAT THE FILE IS GONE.
In VS: Right click on LiveOceanServer (Project), select Python, select Django Synch DB
In the Django Management Console that appears you reply Yes to the prompt to create a new super user.
Create the super user noting the name and password.
This db.sqlite3 file is specific to your local copy of LiveOceanServer; it is not checked in.
If you check it in you clobber the previous version; so don't do that. This deservers a sub-comment on how repositories work.

Repository note; a point of potential confusion
The repository had a copy of db.sqlite3; so that is established as part of the Solution.
I got a local copy that I see in VS.
I deleted my local copy of db.sqlite3. This made no change to the repository.
I created a new db.sqlite3. This has not yet been added to the Solution. It is a file without portfolio.
I may publish my solution to Azure; in which case the db.sqlite3 file goes along and will behave properly: The new super user is present.
This implies that when you click Publish it is the folder contents that are copied to Azure; not moderated by what is in the VS Project.
I may Check In my code to the repository. This will blow away the Repository copy of db.sqlite3 (because I blew it away locally)...
...however! And this is the sinister part: Because I did not add db.sqlite3 to the Solution (the Project) it will not be in the Repository.
(Notice that after it was created it shows up in the LiveOceanServer Project folder but not in VS Solution Explorer.)
If I first right-click LiveOceanServer and Add... new file = db.sqlite3: Now there is a new file that is part of the solution; so a Check In
will copy that into the Repository.
Note: python tools for visual studio is still under active development; so the behavior of Django components may change w/r/t version control
over time. Shifting sands.

Publishing note: File path length error.
This problem is a Windows API limit issue: http://msdn.microsoft.com/en-us/library/aa365247(VS.85).aspx#maxpath
If Publish fails because of a "file length too long" error you may relocate your source tree to a shorter full path.
We created C:\\Src and put everything there using the Workspaces... control in Visual Studio Source Control Explorer.
Note that if you do this there is some "copy rigamarole" necessary to transplant the Python Virtual Environment and
the db.sqlite3 admin file discussed above over to the new location. As follows:
1. DO NOT BLOW AWAY YOUR ORIGINAL COPY OF LIVEOCEANECOSYSTEM; otherwise you've lost your virtual environment.
2. Use Windows File Explorer to locate your original LiveOceanServer project folder and Copy the env sub-folder.
3. Navigate to your new instance of LiveOceanServer and paste the env folder there.
4. Do likewise with db.sqlite3

Step 5. Publish
In VS right click on the LiveOceanServer Project and select Publish.
Here you follow the Wizard; and the key step is Preview. This is the set of litmus tests for whether
the publication will work. The dry run happens on your machine; and if successful it will list all the
files that will be sent to Azure. You then click "Publish".

One possible problem in publishing to Azure is trying to publish without a password. I didn't save the
authentication profile; so one must back up in the Publish wizard to that step, doing it again, and again
checking that the authentication works. Then return to the Publish page of the wizard and proceed.

If the admin pages do not appear to be working... (Kilroy)

Anthony's Added API calls:

api/get-polygon?table=modern&name=Valdez

</pre>
</body>
</html>

whigg / ice2oceans_api Goto Github PK

ice2oceans_api's Introduction

ice2oceans_api's People

Contributors

Watchers

Forkers

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent