updates and improvements to code

compliance_analysis.py

@@ -3,7 +3,14 @@ from utils import set_operator_role_and_location, set_eu_market_status, check_wi
 
 # Create some variables we will use throughout our analysis
 
-dispositive_variables = {
+
+
+# TODO tells the user where the compliance analysis failed
+# TODO cite article from yaml file as explanation
+
+def check_overall_compliance(cards):
+    
+    dispositive_variables = {
     "ai_project_type": {
         "ai_system": False,
         "gpai_model": False,
@@ -24,61 +31,75 @@ dispositive_variables = {
     "data_cc_pass": False,
     "model_cc_pass": False,
     "msg": []
-}
-
-# TODO tells the user where the compliance analysis failed
-# TODO cite article from yaml file as explanation
-
-def check_overall_compliance(dispositive_variables, cc_files):
-
+    }
+    
+    with open(cards['project_file'], 'r') as project_filepath:
+        project_cc = yaml.safe_load(project_filepath.read())
 
+    
     # check intended purposes 
-    # dispositive_variables = check_intended_purpose(dispositive_variables, cc_files)
+    for card in cards['data_file']:
+        dispositive_variables = check_intended_purpose(dispositive_variables, project_cc, card)
+        
+    for card in cards['model_file']:
+        dispositive_variables = check_intended_purpose(dispositive_variables, project_cc, card)
+        
    
     # for each model_cc and data_cc - run analysis with ref to project_cc
     
+    dispositive_variables = run_compliance_analysis_on_project(dispositive_variables, project_cc)
+    
     # dispositive_variables = run_compliance_analysis_on_data(dispositive_variables, data_cc_yaml)
     # dispositive_variables = run_compliance_analysis_on_model(dispositive_variables, model_cc_yaml)
-    
-    dispositive_variables = run_compliance_analysis_on_project(dispositive_variables, project_cc_yaml)
 
     return dispositive_variables
 
-def run_compliance_analysis_on_project(project_cc_yaml): 
+def run_compliance_analysis_on_project(dispositive_variables, project_cc_yaml):
     
     # Project Type    
     if project_cc_yaml['ai_system']['ai_system']['value']:
         dispositive_variables['ai_project_type']['ai_system'] = True
     if project_cc_yaml['gpai_model']['gpai_model']['value']:
         dispositive_variables['ai_project_type']['gpai_model'] = True
-    if dispositive_variables['ai_project_type']['ai_system'] and dispositive_variables['ai_project_type']['gpai_model']:
+    if project_cc_yaml['ai_system']['ai_system']['value'] == True and project_cc_yaml['gpai_model']['gpai_model']['value'] == True:
         dispositive_variables['msg'] = "Your project cannot be both an AI system and a GPAI model. Please revise your Project CC accordingly."
         return dispositive_variables
     
-    if ai_system == True:
-        for key, value in project_cc_yaml['high_risk_ai_system']:
-            if value and sum(map(bool, [project_cc_yaml['high_risk_ai_system']['filter_exception_rights'],project_cc_yaml['high_risk_ai_system']['filter_exception_narrow'],project_cc_yaml['high_risk_ai_system']['filter_exception_human'],project_cc_yaml['high_risk_ai_system']['filter_exception_deviation'], project_cc_yaml['high_risk_ai_system']['filter_exception_prep']])) < 1:
-                project_type = "high_risk_ai_system"
+    # TODO - move file_exceptions to separate section in project_cc?
+    # if dispositive_variables['ai_project_type']['ai_system'] == True:
+    #     for value in project_cc_yaml['high_risk_ai_system']:   
+                     
+    #         if value and sum(map(bool, [
+    #                 project_cc_yaml['high_risk_ai_system']['filter_exception_rights'], 
+    #                 project_cc_yaml['high_risk_ai_system']['filter_exception_narrow'],
+    #                 project_cc_yaml['high_risk_ai_system']['filter_exception_human'],
+    #                 project_cc_yaml['high_risk_ai_system']['filter_exception_deviation'], 
+    #                 project_cc_yaml['high_risk_ai_system']['filter_exception_prep']])
+    #                 ) < 1:
                 
-    if gpai_model == True:
+    #             dispositive_variables['ai_project_type']["high_risk_ai_system"] = True
+               
+    if dispositive_variables['ai_project_type']['gpai_model'] == True:
         if project_cc_yaml['gpai_model_systematic_risk']['evaluation'] or project_cc_yaml['gpai_model_systematic_risk']['flops']:
-            project_type = "gpai_model_systematic_risk"
+            dispositive_variables['ai_project_type']["gpai_model_systematic_risk"] = True
     
     # Operator Type
-    set_operator_role_and_location(dispositive_variables, project_cc_yaml)
-    set_eu_market_status(dispositive_variables, project_cc_yaml)
+    dispositive_variables = set_operator_role_and_location(dispositive_variables, project_cc_yaml)
+    dispositive_variables = set_eu_market_status(dispositive_variables, project_cc_yaml)
 
     # Check if project is within scope of the Compliance Cards project. If not, inform user.
-    if check_within_scope_cc(dispositive_variables):
-        msg = ("Project is within the scope of the Compliance Cards system. Let's continue...") 
+    if check_within_scope_cc(project_cc_yaml):
+        dispositive_variables['msg'].append("Project is within the scope of the Compliance Cards system. Let's continue...") 
     else: 
-        msg = ("Project is not within the scope of the initial version of the Compliance Cards system.")
+        dispositive_variables['msg'].append("Project is not within the scope of the initial version of the Compliance Cards system.")
     
     # Check if the project is within scope of the Act. If it's not, the analysis is over.
-    if check_within_scope_act(dispositive_variables, project_cc_yaml):
-        msg = ("Project is within the scope of Act. Let's continue...") 
+    if check_within_scope_act(project_cc_yaml):
+        dispositive_variables['msg'].append("Project is within the scope of Act. Let's continue...") 
     else: 
-        msg = ("Project is not within the scope of what is regulated by the Act.")
+        dispositive_variables['msg'].append("Project is not within the scope of what is regulated by the Act.")
+
+
 
     # TODO: reactivate the prohibited practices check below 
 
@@ -92,7 +113,7 @@ def run_compliance_analysis_on_project(project_cc_yaml):
 
     # If project is high-risk AI system, check that is has met all the requirements for such systems: 
 
-    if project_type == "high_risk_ai_system":
+    if dispositive_variables['ai_project_type']["high_risk_ai_system"]:
 
     # Do this by examining the Project CC
 
@@ -122,7 +143,7 @@ def run_compliance_analysis_on_project(project_cc_yaml):
     # TODO
     # # If the project is a GPAI model, check that is has met all the requirements for such systems: 
 
-    if gpai_model:
+    if dispositive_variables['ai_project_type']["gpai_model"]:
 
         # # If the project is a GPAI model with systematic risk, check that is has additionally met all the requirements for such systems: 
 
@@ -224,7 +245,7 @@ def run_compliance_analysis_on_model(dispositive_variables, model_cc_yaml):
     
     return dispositive_variables
 
-def check_intended_purpose(dispositive_variables, cc_files):
+def check_intended_purpose(dispositive_variables, project_cc, other_cc):
     
     # We want to run this function for everything classified as a high_risk_ai_system
     # We also need to run it for all 
@@ -245,32 +266,37 @@ def check_intended_purpose(dispositive_variables, cc_files):
     #                     "judicial"]
     
     project_intended_purposes = []
-    for key in project_cc_yaml['high_risk_ai_system']:
-        if project_cc_yaml['high_risk_ai_system'][f'{key}']['value']:
+    for key in project_cc['high_risk_ai_system']:
+        if project_cc['high_risk_ai_system'][f'{key}']['value']:
             project_intended_purposes.append(key) 
     
     # For each Data CC, put the intended uses in a set and then make sure the Project's intended use is in the set
 
     msg = ''
-    dataset_intended_purposes = []
-    for key in data_cc_yaml['high_risk_ai_system']:
-        if data_cc_yaml['high_risk_ai_system'][f'{key}']['value']:
-            dataset_intended_purposes.append(key) 
+    
+    if other_cc['card_type'] == 'data':
+        data_cc = other_cc
+        dataset_intended_purposes = []
+        for key in data_cc['high_risk_ai_system']:
+            if data_cc['high_risk_ai_system'][f'{key}']['value']:
+                dataset_intended_purposes.append(key) 
 
-    for purpose in project_intended_purposes:
-        if purpose not in dataset_intended_purposes:
-            msg = f"You are not compliant because {purpose} is not a valid purpose"
+        for purpose in project_intended_purposes:
+            if purpose not in dataset_intended_purposes:
+                msg = f"You are not compliant because {purpose} is not a valid purpose"
 
     # Now do the exact same thing for all models
 
-    model_intended_purposes = []
-    for key in model_cc_yaml['high_risk_ai_system']:
-        if model_cc_yaml['high_risk_ai_system'][f'{key}']['value']:
-            model_intended_purposes.append(key) 
+    if other_cc['card_type'] == 'model':
+        model_cc = other_cc
+        model_intended_purposes = []
+        for key in model_cc['high_risk_ai_system']:
+            if model_cc['high_risk_ai_system'][f'{key}']['value']:
+                model_intended_purposes.append(key) 
 
-    for purpose in project_intended_purposes:
-        if purpose not in model_intended_purposes:
-            msg = f"You are not compliant because {purpose} is not a valid purpose"
+        for purpose in project_intended_purposes:
+            if purpose not in model_intended_purposes:
+                msg = f"You are not compliant because {purpose} is not a valid purpose"
 
     # TODO return list of intended purpose 
 

project_cc.yaml

@@ -30,13 +30,13 @@ ai_system:
   ai_system: 
     article: 'Art. 3(1)'
     verbose: 'AI project is a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments'
-    value: !!bool true 
+    value: !!bool false 
 
 gpai_model:
   gpai_model: 
     article: 'Art. 3(63)'
     verbose: 'AI project is an AI model, including where such an AI model is trained with a large amount of data using self-supervision at scale, that displays significant generality and is capable of competently performing a wide range of distinct tasks regardless of the way the model is placed on the market and that can be integrated into a variety of downstream systems or applications, except AI models that are used for research, development or prototyping activities before they are placed on the market'
-    value: !!bool true  
+    value: !!bool false  
 
 # Information related to whether or not the project, if an AI system, is a high-risk AI system
 

run.py

@@ -2,7 +2,7 @@ import yaml
 import json
 from pathlib import Path
 import pandas as pd
-from compliance_analysis import run_compliance_analysis_on_project, run_compliance_analysis_on_data, run_compliance_analysis_on_model
+from compliance_analysis import check_overall_compliance
 
 pd.set_option('display.max_columns', None)  
 pd.set_option('display.max_rows', None)
@@ -50,7 +50,6 @@ def gather_cards(files):
     return cards
 
 cards = gather_cards(files)
-print(cards)
 
 # def load_data(files):
 #     cards = []
@@ -58,7 +57,7 @@ print(cards)
 #         with open(file, 'r') as f:
 #             if Path(f.name).name == "project_cc.yaml":
 #                 content = f.read()
-#                 project_cc_yaml = yaml.safe_load(content)
+#                 pcrojet_cc_yaml = yaml.safe_load(content)
 #                 data = project_cc_yaml
 #                 card_type = "project"
 #                 cards.append((card_type, data))
@@ -74,15 +73,11 @@ print(cards)
 #                 cards.append((card_type, data))
 #     return cards
 
-# def process_files(data):
-#     results = []
-#     dispositive_variables = check_overall_compliance(yaml.safe_load(data))
-#     results.append(dispositive_variables['msg'])
-#     return results
+def compliance_analysis(cards):
+    results = []
+    dispositive_variables = check_overall_compliance(cards)
+    results.append(dispositive_variables)#['msg'])
+    return results
 
-# for card in cards:    
-#     data = card[1]
-#     yaml_data = yaml.dump(data, sort_keys=False)
-#     process_files(yaml_data)
-#     print(process_files(yaml_data))
+print(json.dumps(compliance_analysis(cards), indent=4,))
     

utils.py

@@ -6,18 +6,18 @@ import yaml
 def set_operator_role_and_location(dispositive_variables, project_cc_yaml):
     operators = 0
     
-    ai_system = dispositive_variables['ai_project_type']['ai_system']
-    gpai_model = dispositive_variables['ai_project_type']['gpai_model']
-    
-    for var in dispositive_variables['operator_role']:
-        if project_cc_yaml['operator_role'][f'{var}']['value']:
-            dispositive_variables['operator_role'][f'{var}'] = True
+    ai_system = project_cc_yaml['ai_system']['ai_system']['value']
+    gpai_model = project_cc_yaml['gpai_model']['gpai_model']['value']
+
+    for var in project_cc_yaml['operator_details']:
+        if project_cc_yaml['operator_details'][f'{var}']['value'] == True:
+            dispositive_variables['operator_details'][f'{var}'] = True
             operators += 1 
         
     if ai_system and gpai_model:
-        msg = ("Your project cannot be both an AI system and a GPAI model. Please revise your Project CC accordingly.")
+        dispositive_variables['msg'] = ("Your project cannot be both an AI system and a GPAI model. Please revise your Project CC accordingly.")
     if operators != 1:
-        msg = ("Please specify exactly one operator role.")
+        dispositive_variables['msg'] = ("Please specify exactly one operator role.")
     
     return dispositive_variables
 
@@ -28,34 +28,33 @@ def set_eu_market_status(dispositive_variables, project_cc_yaml):
     if project_cc_yaml['eu_market_status']['put_into_service']['value']:
         dispositive_variables['eu_market_status']["put_into_service"] = True
         
-    if project_cc_yaml['operator_role']['output_used']['value']:
-        dispositive_variables['operator_role']["output_used"] = True
+    if project_cc_yaml['operator_details']['output_used']['value']:
+        dispositive_variables['operator_details']["output_used"] = True
         
     return dispositive_variables
 
 
-def check_within_scope_cc(dispositive_variables):
+def check_within_scope_cc(project_cc_yaml):
 
     # Check that the person filling out the form (the operator) is in fact a provider;
-    
-    if dispositive_variables['operator_details']['provider']:
+    if project_cc_yaml['operator_details']['provider']['value']:        
         return True
     else:
         print("The initial versiton of the Compliance Cards System is for provider-side compliance analyses only.")
         return False
 
-def check_within_scope_act(dispositive_variables, project_cc_yaml):
+def check_within_scope_act(project_cc_yaml):
 
     # Check that the project is within the scope of the Act 
     
-    ai_system = dispositive_variables['ai_project_type']['ai_system']
-    gpai_model = dispositive_variables['ai_project_type']['gpai_model']
+    ai_system = project_cc_yaml['ai_system']['ai_system']
+    gpai_model = project_cc_yaml['gpai_model']['gpai_model']
 
-    placed_on_market = dispositive_variables['eu_market_status']['placed_on_market']
-    put_into_service = dispositive_variables['eu_market_status']['put_into_service']
+    placed_on_market = project_cc_yaml['eu_market_status']['placed_on_market']
+    put_into_service = project_cc_yaml['eu_market_status']['put_into_service']
     
-    eu_located = dispositive_variables['operator_details']['eu_located']
-    output_used = dispositive_variables['operator_details']['output_used']
+    eu_located = project_cc_yaml['operator_details']['eu_located']
+    output_used = project_cc_yaml['operator_details']['output_used']
     
     if not check_excepted(project_cc_yaml):
         if ((ai_system and (placed_on_market or put_into_service)) or (gpai_model and placed_on_market)):   # Article 2.1(a)
@@ -74,7 +73,7 @@ def check_excepted(project_cc_yaml):
         return False 
 
 # TODO update function
-def check_prohibited(dispositive_variables, project_cc_yaml):
+def check_prohibited(project_cc_yaml):
 
     ai_system = project_variables['ai_project_type']['ai_system']