generated genesis should be OK

lib/functions.py

-import json
+import json, csv
 from symbol import parameters
+import sys
 from urllib.request import urlopen
 from substrateinterface import Keypair, KeypairType
 from currency_parameters import *
+from bs4 import BeautifulSoup
 
-def get_input_data():
+def get_g1stats_data():
     f = urlopen("https://g1-stats.axiom-team.fr/data/forbes.json")
     return json.load(f)
 
+def get_datajune_certs():
+    list_certs = {}
+    certsIndex = urlopen("https://files.datajune.coinduf.eu/graphs.lg").read().decode()
+    
+    soup = BeautifulSoup(certsIndex, features="lxml")
+    for chunks in soup.find_all('a'):
+        pass
+    lastChunk = int(chunks.string.split('.')[0])
+    
+    print('Getting certifications in progress, this could take ~5 minutes...')
+    
+    for i in progressbar(range(1, lastChunk)):
+        # print(i)
+        paddedNbr = str(i).rjust(4, '0')
+        graphData = urlopen("https://files.datajune.coinduf.eu/graphs.lg/" + paddedNbr + ".lg")
+        
+        for j in graphData.readlines():
+            data = j.decode()
+            if 'simplegraph' in data: continue
+            
+            cert_data = data.replace('\n', '').split(',')
+            issuer = int(cert_data[0])
+            receiver = int(cert_data[1])
+            if issuer not in list_certs: list_certs.update({issuer: []})
+            if receiver not in list_certs[issuer]: list_certs[issuer].append(receiver)
+            # if i == 20: break
+    
+    list_certs_json = json.dumps(list_certs).encode()
+    list_certs_json_file = open('certs.json', 'wb')
+    list_certs_json_file.write(list_certs_json)
+    # print(list_certs[1])
+    # sys.exit(0)
+    
+    return list_certs
+
+def get_datajune_certs_local():
+    f = open('certs.json')
+    return json.load(f)
+        
+def username_to_number(username, datajune_usernames):
+    if username in datajune_usernames:
+        return datajune_usernames.index(username) + 1
+    else:
+        return 9999999
+
+def number_to_username(number, datajune_usernames):
+    return datajune_usernames[number - 1]
+
+
+def get_datajune_usernames():
+    return urlopen("https://files.datajune.coinduf.eu/global/pseudos.txt").read().decode().split()
+
+
+
 def v1PubkeyToV2Address(pubkey):
     # get incomplete Substrate keypair (only public key) from public key bytes
     keypair = Keypair(public_key=pubkey, ss58_format=42, crypto_type=KeypairType.ED25519)
@@ -64,3 +120,17 @@ def get_technical_committee():
     return ['poka', 'elois', 'HugoTrentesaux' 'vit', 'tuxmain']
 
 
+
+
+import sys
+def progressbar(it, prefix="", size=60, out=sys.stdout): # Python3.3+
+    count = len(it)
+    def show(j):
+        x = int(size*j/count)
+        print("{}[{}{}] {}/{}".format(prefix, "#"*x, "."*(size-x), j, count), 
+                end='\r', file=out, flush=True)
+    show(0)
+    for i, item in enumerate(it):
+        yield item
+        show(i+1)
+    print("\n", flush=True, file=out)
\ No newline at end of file

main.py

@@ -7,8 +7,11 @@ from currency_parameters import *
 # Final ĞTest genesis JSON
 gtest_genesis = {}
 
-# Get input data from g1-stats
-inputData = get_input_data()
+# Get inputs data
+g1stats_data = get_g1stats_data()
+datajune_usernames = get_datajune_usernames()
+# datajune_certs = get_datajune_certs()
+datajune_certs = get_datajune_certs_local()
 
 # Get ĞTest parameters
 genesis_parameters = get_genesis_parameters()
@@ -29,7 +32,7 @@ gtest_genesis.update({'sudo_key': sudo_key})
 # Add identities
 gtest_genesis.update({'identities': {}})
 i=0
-for profile in inputData:
+for profile in g1stats_data:
     i=i+1
     
     # skip 11111111111111111111111111111111111111111111 invalid pubkey, don't know why
@@ -50,10 +53,35 @@ for profile in inputData:
     # Get username if exist
     if profile['id'] != None:
         username = profile['id']['username']
+        # if username != 'poka': continue
+        
+        # Get certs for this identity
+        # sent_certs = set()
+        username_number = username_to_number(username, datajune_usernames)
+        
+        # print(username)
+        # for j in datajune_certs:
+        #     cert_data = j.split(',')
+        #     issuer = int(cert_data[0])
+        #     receiver = int(cert_data[1])
+            
+        #     if issuer == username_number:
+        #         receiver_username = number_to_username(receiver, datajune_usernames)
+        #         sent_certs.add(receiver_username)
+        #         # print(receiver_username)
+        
+        sent_certs = []
+        if str(username_number) not in datajune_certs:
+            print(username + " doesn't exist !")
+            continue
+        for idty in datajune_certs[str(username_number)]:
+            username_receiver = number_to_username(idty, datajune_usernames)
+            sent_certs.append(username_receiver)
+        
         # Format identities genesis JSON
-        gtest_genesis['identities'].update({username: {'balance': balance,'certs': [],'pubkey': address_v2}})
+        # print(sent_certs)
+        gtest_genesis['identities'].update({username: {'balance': balance,'certs': sent_certs,'pubkey': address_v2}})
         
-    # TODO: get certs by identities
 
 # Transform data to JSON
 gtest_genesis_json = json.dumps(gtest_genesis, indent=2).encode()