pyshopify 0.9.4

Shopify Order API library with containeriezed MS SQL database

Shopify Orders Rest API Wrapper and Data Export

The purpose of this repository is to allow easier Shopify data analysis using any SQL or and BI tools that support sql. Shopify's rest API is not data analysis friendly, returning one large, denormalized dataset. This library yields a more normalized dataset that can be used to populate a database or perform a direct analysis.

This repository has two part -

1. Python app - pulls data from shopify orders api into dataframes
2. Docker container - fully self contained database and app

This has the flexibility of exporting the data to a SQL Server ( along with other relational dbs), exporting to csv and returning a dictionary of dataframes.

The Structure of the SQL Database is described in depth in the Docs folder, an SQL script to build the database is in the scripts folder setup.sql.

A fully contained docker container with docker-compose.yml is included for simple deployment. Container contains Microsoft SQL Server 2019 with this library installed for easily running with a small amount of configuration.

Table of Contents

1. Configuration
2. Base Installation
3. Installation with SQL Drivers
4. Running the Application as a Module
5. Running as Command Line Applicaiton
6. Database Structure
   • Tables
   • Stored Procedures
   • Full Documentation
7. Database Script
8. Docker Container
9. Useful Queries

Script Configuration file

The python script is configured through an ini based file. Here is the template:

[shopify]
# Orders API Endpoint
url_base = https://**YOURSTORE**.myshopify.com
order_ep = /admin/api/2020-10/orders.json

# Shopify Access Token
access_token = ***ACCESS TOKEN FROM ADMIN***

# Earliest date
early_date = 20000101

# Items per Return - 250 Max
items_per_page = 250

# Uncomment to get data between two dates
# Overrides past days of history
# start = 20210610
# end = 20210617

# Get past days of history - 30 days is default
days = 30

[sql]
enable = False

# Driver for pyodbc to use
driver = ODBC Driver 17 for SQL Server

# Database server & port
server = ***DATABASE SERVER***
port = 1433
database = ***DATABASE***

# Database user & password
db_user = ***DB USER***
db_pass = ***DB PASSWORD***

[csv]
enable = True
# Relative filepath of csv folder output
filepath = csv_export

[custom]
#Output dictionary of dataframes
enable = False

The configuration file will default to config.ini in the current working directory. A custom path can be defined by passing to ShopifyApp('rel/path/config.ini') when instantiating or the shopify_cli --config rel/path/config.ini

Installing Base Library

Install from PyPi through pip install. This will install all required dependencies for running to export to CSV and return a dictionary of dataframes.

$ python3 -m pip install pyshopify

Library with SQL Driver

In order to use the sql output feature, the database driver and python library must be installed.

On linux it takes several steps. pyodbc must be installed as root user.

$ apt-get install -y unixodbc-dev msodbcsql17
$ sudo -H python3 -m pip install pyodbc
$ python3 -m pip install sqlalchemy 

Ensure MS ODBC driver is installed on Windows. Can be found Microsoft SQL OBC Driver

Running the ShopifyApp()

The primary class is ShopifyApp() which contains all of the necessary execution methods.

Can be run two ways:

As a module, instantiating ShopifyApp and running app_runner(). If custom is disabled, it will return None. If custom is enabled, a dictionary of dataframes of the entire api call is returned.

from pyshopify.runner import ShopifyApp

# If no value is passed, the program will try to use a file in the working directory named config.ini
shop_instance = ShopifyApp('rel/dir/to/config.ini')

run = shop_instance.app_runner()

By enabling custom in the configuration, the entire API call is combined into a single dictionary. This can be a large amount of data, and want to apply logic in between each call. In this case, use the app_iterator() method

api_iterator = shop_instance.app_iterator()
for api_return in api_iterator:
    #Perform logic here, return dictionary is in same structure as the full dictionary from app_runner()
    print(api_iterator.get('Refunds'))

The configuration can be set after you've instantiated the class:

shop_instance = ShopifyApp()
shop_instance.csv_enable = True # enables csv output
shop_instance.custom_enable = True # Enabled full output

Enabling SQL output is possible but slightly more involved.

shop_instance = ShopifyApp()
shop_instance.sql_conf = {
    'enabled': True,
    'driver': 'ODBC Driver 17 for SQL Server',
    'server': 'localhost',
    'database': 'shop_rest',
    'db_user': 'sa',
    'db_password': 'StrongDBPassword'
}
shop_app.sql_alive = shop_app.sql_connect()

shop_app.sql_alive should return true if connected to database. An error will be raised if unable to connect

The return value of app_runner() and app_iterator() is a dictionary of dataframes in the structure below. This structure mirrors the structure of the database output.

from pyshopify.runner import ShopifyApp
shop_class = ShopifyApp()
run = ShopifyApp.app_runner()

# Order Details
# ['id' 'order_date' 'fulfillment_status' 'name' 'number' 'order_number', 'payment_gateway_names' 'processing_method' 'source_name', 'subtotal_price' 'total_discounts' 'total_line_items_price' 'total_price', 'total_price_usd' 'total_tax' 'total_weight', 'email', 'updated_at']
orders_dataframe = run.get("Orders")

# Refunds Dataframe with Date of Refund and Order ID
# Columns    ['refund_date', 'order_id']
refunds_dataframe = run.get("Refunds")

# Refund Line Items Showing Units returned
# ['id', 'line_item_id', 'quantity', 'subtotal', 'total_tax', 'variant_id', 'refund_id', 'order_id']
refund_lineitems = run.get("RefundLineItems")

# Line Items sold
# ['id', 'order_id', 'order_date', 'variant_id', 'quantity', 'price']
line_items = run.get("LineItems")

# Customer info for each order
# ['order_id', 'order_date', 'email', 'customer_id', 'orders_count', 'total_spent', 'created_at']
customer_orders = run.get("OrderCustomers")

# Order Shipping Prices
# [id, order_id, order_date, carrier_identifier, code, delivery_category, ship_discount_price, ship_price, phone, requested_fulfillment_id, source, title]
shipping_lines = run.get("ShipLines")

# Discount Applications
# ['order_id', 'order_date', 'type', 'code', 'title', 'description', 'value', 'value_type', 'allocation_method', 'target_selection', 'target_type']
discount_applications = run.get("DiscountApps")

# Discount Codes
# ['order_id', 'code', 'type', 'amount']
discount_codes = run.get('DiscountCodes')
 

Running as Command Line Application

From command line:

# There are different command options
$ shopify_cli --help
Usage: shopify_cli [OPTIONS]

  Run Shopify App CLI.

Options:
  -d, --days INTEGER     get days of history
  -b, --between TEXT...  get between 2 dates - yyyy-MM-dd, ex -b 2020-01-01
                         2020-01-02
  --sql-out / --no-sql   write to database - Default False
  --csv-out / --no-csv   Write results to csv files - Default true
  --csv-location TEXT    Relative location of csv export folder defaults to
                         csv_export/
  --config TEXT          Relative location of config.ini - defaults to
                         config.ini in currect directory
  --help                 Show this message and exit.

# Default arguments are:
$ shopify_cli -d 30 --csv-out --no-sql --config config.ini

$ shopify_cli -d 30 # get last 30 days of orders and export to CSV in CWD/csv_export
$ shopify_cli -b 2020-01-01 2020-01-02 # get orders between dates and export csv
$ shopify_cli -d 30 --sql-out # update SQL db

Enabling the custom section of config.ini will return a dictionary of order data parsed into separate key-value pairs that allow easier processing and analysis. Using the sql option will update an SQL Database with the data from the dictionary.

The structure of the custom return dictionary reflects the SQL database structure that it will update:

from pyshopify import ShopifyApp

# Enable custom in config.ini
shop_app = ShopifyApp()

app_run = shop_app.app_runner()

app_run = {
      'Orders': OrdersDataframe,
      'Customers': CustomersDataframe,
      'LineItems': LineItemDataFrame
      'Refunds': RefundDataFrame,
      'RefundLineItem': RefLineItem.DataFrame,
      'Adjustments': RefundAdjustmentsDataFrame
}

Each dataframe in the return represents an SQL Table in the database. The dataframe column types match the database column types.

Database Structure

Exporting SQL from API response is two step process:

1. Send DataFrame to temporary SQL Table
2. Run stored procedure to merge temp table with the appropriate table

The full database documentation is located here

Click on each item for more details.

shop_rest is the default database name. Remember to change config.ini if using a different database name.

Tables

Name	Description
dbo.Adjustments	Order Refund Adjustments
dbo.DateDimension	Date Dimension Table for Analysis
dbo.DiscountApps	Discount Applications for Each Order
dbo.DiscountCodes	Discount Codes In Use for Each Order
dbo.LineItems	Line Items with Units Sold for Orders
dbo.OrderCustomers	Customer Info based on Order ID
dbo.Orders	Order Details
dbo.RefundLineItem	Refunded Units
dbo.Refunds	Order Refunds
dbo.ShipLines	Order shipping lines

Stored Procedures

Name	Description
dbo.adjustments_update	Update Adjustments
dbo.cust_update	Update Customer Orders Table
dbo.discapp_update	Update discounts applied to each order
dbo.disccode_update	Update discount codes used for each order
dbo.lineitems_update	Update Line Items
dbo.orders_update	Merge Orders
dbo.reflineitem_update	Merge Refunded Line Items
dbo.refunds_update	Merge Refunds
dbo.shipline_update	Update shipping lines table

A DateDimension table is included for easier analysis

Database script

To build the dataabse, run the setup.sql script in the docker/scripts folder. This has only been tested on Microsoft SQL Server 2019 but can easily be adapted for other databases. It will set up all of the required tables and stored procedures.

The DateDimension table can be created from the dates.sql script in the docker/scripts fold. Credit goes to Aaron Bertrand for this amazing script - Creating a date dimension or calendar table in SQL Server

Both are ran automatically when starting the docker container.

Docker Container

There is a Dockerfile and docker-compose.yml in the docker folder. This is based on the Microsoft SQL Server 2019 container running on Ubuntu. It installs all of the necessary applications to run pyshopify and a database instance to write to.

NOTE: This is NOT production ready. Security is not hardened, container is run as root user.

Both containers automatically deploy setup.sql and dates.sql to build the required database structure.

Please make sure to set the password in the docker-compose.yml file.

Download entire docker folder or just clone repo

$ git clone https://github.com/webdjoe/pyshopify
$ cd docker

Use vim or nano to edit docker-compose.yml and config.ini in config folder

$ vim docker-compose.yml
$ vim config/config.ini

Build & Run container. Use -d for detached

$ docker-compose build
$ docker-compose run -d

Once started test if SQL server is running. A 0 return value indicates the server has started up.

$ docker exec -it shopsql /opt/mssql-tools/bin/sqlcmd -h -1 -t 1 -U sa -P "$SA_PASSWORD" -Q "SET NOCOUNT ON; Select SUM(state) from sys.databases")

shopify_cli can be called in container to update the container's database.

Get last 30 days of data and import into SQL Server running in container
$ docker exec -it shopsql shopify_cli -d 30 --no-csv --sql-out

Get data between dates and import into SQL Server running on container 
$ docker exec -it shopsql shopify_cli -b 20200101 20201231 --no-csv --sql-out

shopify_cli can also be run from container to get data in csv files in the csv_export folder Assumes run from the same folder as the docker files. Container links csv_export folder to retreive export without having to copy from container.

$ docker exec -it shopsql shopify_cli --csv-out
$ cd csv_export

Useful Queries

These scripts assume Shopify has been configured correctly from the start, meaning SKU's, product names and product titles remained consistent. If not, so data cleansing may be required. For example, if your product title has been properly configured you can update SKU's from the variant_sku column.

List all product names, titles, skus and variant SKU's to check for consistency.

Resulting table is ordered by quantity sold to check impact of product sku & naming consistency

SELECT [variant_id], [name], [product_id], [sku], [title], SUM([quantity]) quantity
FROM shop_rest.dbo.LineItems
GROUP BY [variant_id], [name], [product_id], [sku], [title]
ORDER BY SUM([quantity])

Or check if above a minimum threshold, 10 in this case:

SELECT [variant_id], [name], [product_id], [sku], [title], SUM([quantity]) quantity
FROM shop_rest.dbo.LineItems
GROUP BY [variant_id], [name], [product_id], [sku], [title]
HAVING SUM([quantity]) > 10
ORDER BY SUM([quantity])

If there are inconsistencies, an easy way to clear up is to import the return table into excel and make a table of variant_id with the correct title and SKU. Import that table into SQL server and join with queries that return variant_id.

Quantities Sold

Return quantities sold, grouped by month. SKU and title are used as identifiers as identifiers.

SELECT EOMONTH([order_date],0) OrderDate, [sku], [name], SUM([quantity]) Quantity
FROM shop_rest.dbo.LineItems
GROUP BY EOMONTH([order_date], 0), [sku], [name]
ORDER BY [order_date]

If there is inconsistency in product naming and SKU's, use a subquery to join table with the table created containing the variant_id with the corrected SKU's and product titles. The following logic can be applied to any other query containing variant_id

SELECT te.OrderDate, te.sku, SUM(te.qty) Quantity FROM

(
    SELECT CAST(EOMONTH([order_date],0) as date) OrderDate
        ,vs.[sku]
        ,vs.[name]
        ,[sku]
        ,li.[quantity] qty
    FROM shop_rest.dbo.LineItems li
    INNER JOIN shop_rest.dbo.VariantSkus vs 
        ON li.variant_id = vs.variant_id
) te

GROUP BY te.OrderDate, te.sku
ORDER BY te.OrderDate

Number of returning customers per month

SELECT COUNT(*) AS Customers, CAST(EOMONTH(order_date,0) AS date) AS OrderDate
    FROM dbo.OrderCustomers
WHERE [orders_count] > 1
GROUP BY CAST(EOMONTH(order_date) AS date)

Monthly Gross Sales with Discounts.

SELECT CAST(EOMONTH([order_date],0) AS date) Month
      ,SUM([total_line_items_price]) GrossPrice
	  ,SUM([total_discounts]) TotalDiscount
      ,SUM([subtotal_price]) Subtotal
  FROM [shop_rest].[dbo].[Orders]  
  GROUP BY CAST(EOMONTH([order_date],0) AS DATE)

Monthly Sales, Discounts & Shipping Revenue

SELECT 
     ord.[Month]
    ,ord.[GrossPrice]
    ,ord.[TotalDiscount]
    ,ord.[Subtotal]
    ,ship.[ShipPrice]
    ,(ord.[GrossPrice] - ord.[TotalDiscount] + ship.[ShipPrice]) TotalPrice
FROM (
    SELECT CAST(EOMONTH([order_date],0) AS date) Month
          ,SUM([total_line_items_price]) GrossPrice
          ,SUM([total_discounts]) TotalDiscount
          ,SUM([subtotal_price]) Subtotal
          ,SUM([total_tax]) TotalTax
          ,SUM([total_price]) TotalPrice
    FROM [shop_rest].[dbo].[Orders]  
    GROUP BY CAST(EOMONTH([order_date],0) AS DATE)
) ord 
INNER JOIN (
    SELECT SUM([ship_price]) ShipPrice, CAST(EOMONTH([order_date],0) AS date) Month
    FROM [shop_rest].[dbo].[ShipLines]
    GROUP BY CAST(EOMONTH([order_date],0) AS date)
) ship
ON ord.[Month] = ship.[Month]
  ORDER BY ord.[Month]