Merge pull request #7 from TechForPalestine/unify-brands

Recenter yaml files around brands instead of companies

.github/workflows/ci.yml

@@ -22,23 +22,17 @@ jobs:
 
     # Steps represent a sequence of tasks that will be executed as part of the job
     steps:
-      # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
-      - uses: actions/checkout@v3
-
-      # Runs a single command using the runners shell
-      - name: Run a one-line script
-        run: echo Hello, world!
-
       # Setup and run python script, generates a json file
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v5
         with:
           python-version: '3.12'
           cache: 'pip' # caching pip dependencies
-      - run: |
-          pip install -r requirements.txt
+      - run: pip install -r requirements.txt
+      - name: Validate YAML
+        run: |
           python --version
-          python validate_model.py
+          python validate_yaml.py
 
   if_merged:
     if: github.event.pull_request.merged == true
@@ -55,7 +49,7 @@ jobs:
           cache: 'pip' # caching pip dependencies
       - run: |
           pip install -r requirements.txt
-          python validate_model.py
+          python validate_yaml.py
           python generate_model.py
       - uses: mikeal/publish-to-github-action@master
         env:

.gitignore

@@ -1 +1,2 @@
 .DS_Store
+.rtx.toml

README.md

@@ -11,50 +11,13 @@ Sources:
 
 - https://pastebin.com/raw/ks9GRE4L
 
+## Data
+All data is inputted & stored as YAML files in the `data/` directory.
+Output formats, such as CSV and JSON are in the `generated/` directory.
 
-## Model
-
-Standardised data model for each company:
-
-Example: 
-
-```yaml
-
-  # name: required, brand name
-- name: "Brand"
-
-  # website: optional, website of company, useful for alternatives to help users navigate to and shop from
-  website:
-  # image_url: optional, link to brand logo image
-  image_url: |-
-    https://1000logos.net/wp-content/uploads/2023/03/Whiskas-Logo-2003.png
-
-  # categories/tags: optional, for grouping data, makes it easier to find alternatives
-  categories: []
-
-  # parents: required, list of parents, since one brand can have different parent companies in a different country/region
-  parents:
-    # name: required, parent company name
-  - name: "Parent Company"
-
-    # location: required, list of locations where this is the parent company of the brand
-    location: [global]
-
-    # level: required, boycotting level, one of direct, alternative
-    level: direct		
-
-    # details: optional - boycotting details, relevant for boycott brand, can be empty for alternative
-    #   reason: reason for boycotting (if applicable)
-    #   source_url: evidence for why this brand/parent should be boycotted
-    details:			
-      reason: |-
-        **Wilson Partnership**
-
-        **Wilson** has partnered with **Delta Galil Industries**, Ltd. (DELT/Tel Aviv Stock Exchange), the global manufacturer and marketer of branded and private label apparel products for men, women, and children. **Delta Galil Industries** is an **Israeli** textile firm headquartered in **Tel Aviv**, with plants around the world.
-      source_url: |-
-        https://deltagalil.com/brands/licensed-brands/default.aspx
-```
-
+Schemas for the YAML data can be found in the `schemas` directory, along with descriptions for each field.
+These schemas are in [JSON Schema](https://json-schema.org/) format, but represented in YAML for simplicity.
+The `validate_yaml.py` script validates all brands and companies using the schemas.
 
 ## Useful Resources & Links
 

data/companies/ahava.yaml → data/brands/ahava.yaml

@@ -1,12 +1,13 @@
 name: "Ahava"
-reason: |-
+status: avoid
+reasons: [operations_in_settlements, operations_in_israel]
+countries: [global]
+website: https://www.ahava.com/
+logo_url: https://upload.wikimedia.org/wikipedia/commons/8/81/AhavaLogo.png
+description: |-
   Ahava is an Israeli cosmetics company which operates its main manufacturing plant and showroom in Mitzpe Shalem, 
   an illegal settlement in the West Bank.[^1]
   The Palestinian BDS National Committee has called for a boycott of Ahava.[^2]
   
   [^1]: https://en.wikipedia.org/wiki/Ahava
-  [^2]: https://bdsmovement.net/Act-Now-Against-These-Companies-Profiting-From-Genocide
-brands:
-  - name: Ahava
-    slug: ahava
-    alternatives: []
+  [^2]: https://bdsmovement.net/Act-Now-Against-These-Companies-Profiting-From-Genocide

data/brands/drinkmate.yaml

@@ -0,0 +1,5 @@
+name: "Drinkmate"
+status: neutral
+countries: [us, ca, gb, eu, au, co, pe, ae, sa, jp, sg, tw]
+website: https://idrinkproducts.com/
+description: Drinkmate offers sparkling water and soda makers (also known as soda machines).

data/companies/puma-se.yaml → data/brands/puma.yaml

@@ -1,11 +1,10 @@
-name: "Puma SE"
-reason: |-
+name: "Puma"
+status: avoid
+reasons: [operations_in_israel]
+countries: [global]
+description: |-
   The Palestinian BDS National Committee has called for a boycott of Puma[^1] for several reasons, including their 
   sponsorship of the Israeli Football Association (IFA), which includes teams in illegal settlements on occupied 
   Palestinian land.
   
   [^1]: https://bdsmovement.net/boycott-puma
-brands:
-  - name: Puma
-    slug: puma
-    alternatives: []

data/brands/sabra.yaml

@@ -0,0 +1,16 @@
+name: "Sabra"
+status: avoid
+reasons: [operations_in_israel]
+categories: [food]
+website: https://sabra.com/
+description: |-
+  Sabra was an Isareli hummus company, and is now a joint venture between Pepsi and Strauss.  Strauss is an Israeli food company.
+alternatives_text: |-
+  Most hummus is not Israeli.  Any non-Israeli hummus at your local grocery store is preferred over Sabra.
+stakeholders:
+  - id: pepsico
+    type: owner
+    percent: 50
+  - id: strauss-group
+    type: owner
+    percent: 50

data/brands/sodastream.yaml

@@ -0,0 +1,18 @@
+name: "SodaStream"
+status: avoid
+reasons: [operations_in_israel]
+countries: [global]
+website: https://sodastream.com/
+logo_url: https://upload.wikimedia.org/wikipedia/commons/thumb/6/63/Soda_Stream_Logo.svg/2560px-Soda_Stream_Logo.svg.png
+description: |- # optional
+  SodaStream has long been a BDS target due to their operation of factories on stolen land and their racial discrimination against Palestinian workers.[^1]
+  
+  [^1]: https://bdsmovement.net/Act-Now-Against-These-Companies-Profiting-From-Genocide
+alternatives: [drinkmate]
+stakeholders:
+  - id: pepsi
+    type: owner
+    percent: 50
+  - id: strauss-group
+    type: owner
+    percent: 50

data/companies/pepsico.yaml

@@ -1,16 +1,10 @@
 name: "PepsiCo, Inc."
-reason: |-
+status: avoid
+description: |-
   PepsiCo purchased the Israeli company SodaStream in August 2018.[^1] SodaStream has long been a BDS target due to 
   their operation of factories on stolen land and their racial discrimination against Palestinian workers.[^2]
 
   Pepsi also owns 50% of Sabra, with the other 50% being owned by Israeli food company Strauss.
   
   [^1]: https://www.nasdaq.com/articles/sodastream-sells-out-price-too-low-2018-08-20
   [^2]: https://bdsmovement.net/Act-Now-Against-These-Companies-Profiting-From-Genocide
-brands:
-  - name: SodaStream
-    slug: sodastream
-    alternatives: [drinkmate]
-  - name: Sabra
-    slug: sabra
-    alternatives: []

data/companies/strauss-group.yaml

@@ -0,0 +1,4 @@
+name: "Strauss Group Ltd."
+status: avoid
+description: |-
+  Strauss Group is an Israeli manufacturer and marketer of consumer foods.

requirements.txt

@@ -1 +1,2 @@
-
+jsonschema == 4.20.0
+pyyaml == 6.0.1

schemas/brand_schema.yaml

@@ -0,0 +1,92 @@
+type: object
+description: A brand is something that a consumer may search to know whether or not to purchase a product.
+properties:
+  name:
+    description: Name of the brand
+    type: string
+  description:
+    description: Explanation of why the brand is on this list.  Or, if it's not a brand to avoid, just a description of the brand.
+    type: string
+  status:
+    description: Whether to support or avoid this brand
+    type: string
+    enum:
+      - support
+      - neutral
+      - avoid
+  reasons:
+    description: A list of reasons for why consumers should avoid this brand.
+    type: array
+    items:
+      type: string
+      enum:
+        - operations_in_israel
+        - operations_in_settlements
+  countries:
+    description: |-
+      A list of countries (ISO alpha-2 country codes) that the brand operates in. 
+      Useful for filtering for brands that are relevant to a specific region.
+      "global" means it is available in all countries.
+    type: array
+    items:
+      anyOf:
+        - type: string
+          enum: [global]
+        - type: string
+          pattern: "^[a-z]{2}$"
+  categories:
+    type: array
+    items:
+      type: string
+      enum: [food]
+  website:
+    type: string
+    format: uri
+    pattern: "^https?://"
+  logo_url:
+    description: Logo should be at least 200x200 pixels
+    type: string
+    format: uri
+    pattern: "^https?://"
+  alternatives:
+    description: List of brands that would be an alternative option to purchasing from this brand.
+    type: array
+    items:
+      type: string
+      pattern: "^[a-z][a-z-]+[a-z]$"
+  alternatives_text:
+    description: |-
+      Plain text description of alternatives.
+      Especially useful when alternatives are difficult to enumerate.
+    type: string
+  stakeholders:
+    description: |-
+      If useful, a list of stakeholders, such as companies who own this brand.
+    type: array
+    items:
+      anyOf:
+        - type: object
+          properties:
+            id:
+              type: string
+              pattern: "^[a-z][a-z-]+[a-z]$"
+            type:
+              type: string
+              enum: [owner]
+            ownership_percent:
+              type: number
+              minimum: 10
+              maximum: 100
+  allOf:
+    if:
+      properties:
+        status:
+          const: avoid
+      required: [status]
+    then:
+      required: [reasons]
+required:
+  - name
+  - status
+  - description
+additionalProperties: false

schemas/company_schema.yaml

@@ -0,0 +1,21 @@
+type: object
+description: A company is an entity that owns more than one brand, or owns other companies.
+properties:
+  name:
+    type: string
+    description: Name of the company
+  description:
+    type: string
+    description: Description of the company, including why it is on the list
+  status:
+    type: string
+    description: Whether to support or avoid this company
+    enum:
+      - support
+      - neutral
+      - avoid
+required:
+  - name
+  - status
+  - description
+additionalProperties: false

validate_yaml.py

@@ -0,0 +1,47 @@
+import os
+import glob
+import yaml
+import logging
+import jsonschema
+from jsonschema import validate
+
+def get_filename_only(file_path):
+  base_name = os.path.basename(file_path)
+  filename_only, _ = os.path.splitext(base_name)
+  return filename_only
+
+def load_yaml(file_path):
+  with open(file_path, 'r') as file:
+    return yaml.safe_load(file)
+
+def validate_with_schema(file_path, schema):
+  try:
+    validate(load_yaml(file_path), schema)
+    return True
+  except jsonschema.exceptions.ValidationError as ve:
+    logging.error('Validation error in ' + get_filename_only(file_path))
+    logging.error(ve)
+    return False
+
+def main():
+  global root_path
+  root_path = os.path.dirname(os.path.realpath(__file__))
+
+  brand_schema = load_yaml(os.path.join(root_path, 'schemas/brand_schema.yaml'))
+  brand_files = glob.glob(os.path.join(root_path, 'data/brands/') + '*.yaml')
+  print('Validating', len(brand_files), 'brands')
+  for file in brand_files:
+    if not validate_with_schema(file, brand_schema):
+      exit(1)
+  print('All brands are valid.')
+
+  company_schema = load_yaml(os.path.join(root_path, 'schemas/company_schema.yaml'))
+  company_files = glob.glob(os.path.join(root_path, 'data/companies/') + '*.yaml')
+  print('Validating', len(company_files), 'companies')
+  for file in company_files:
+    if not validate_with_schema(file, company_schema):
+      exit(1)
+  print('All companies are valid.')
+
+if __name__ == "__main__":
+  main()