Merge branch 'main' of github.com:Fraunhofer-AISEC/cpg-contract-check…

…er into main

data-collection/snippets/parser/analyze_compiler_versions.py

@@ -0,0 +1,137 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+import os
+import re
+import time
+import pprint
+import pymongo
+import datetime
+
+from web3 import Web3
+from tqdm import tqdm
+from packaging import version
+
+MONGO_HOST = "sectrs-vascodagama"
+MONGO_PORT = 27017
+
+def extract_version(source_code):
+    versions = re.compile("pragma solidity (.+?);").findall(source_code)
+    if len(versions) == 0:
+        return None
+    v = " ".join(versions)
+    v = v.replace('"', '')
+    v = v.replace("\t", "")
+    v = v.replace(" .", ".")
+    v = v.replace(". ", ".")
+    v = v.replace("||", " ")
+    v = v.replace("-", " ")
+    v = v.replace("x", "0")
+    v = v.replace("*", "0")
+    v = v.replace("^", "")
+    v = v.replace("~", "")
+    v = v.replace("=", " ")
+    v = v.replace(">", " ")
+    v = v.replace("<", " ")
+    v = v.replace("=", " ")
+    v = " ".join(v.split())
+    v = v.strip()
+    if len(v.split(" ")) != 1:
+        min = None
+        for w in v.split(" "):
+            if min == None or version.parse(w) < version.parse(min):
+                min = w
+        v = min
+    v = v.replace(".99", ".9")
+    v = v.replace(".00", ".0")
+    if len(v.split(".")) == 4:
+        v = v.replace(".0", "")
+    if v.endswith(".01"):
+        v = v.replace(".01", ".1")
+    if v.endswith(".02"):
+        v = v.replace(".02", ".2")
+    if v.endswith(".03"):
+        v = v.replace(".03", ".3")
+    if v.endswith(".04"):
+        v = v.replace(".04", ".4")
+    if v.endswith(".05"):
+        v = v.replace(".05", ".5")
+    if v.endswith(".06"):
+        v = v.replace(".06", ".6")
+    if v.endswith(".07"):
+        v = v.replace(".07", ".7")
+    if v.endswith(".08"):
+        v = v.replace(".08", ".8")
+    if v.endswith(".09"):
+        v = v.replace(".09", ".9")
+    if v.startswith("8.") and v.endswith(".0"):
+        v = "0."+v.replace(".0", "")
+    if len(v.split(".")) == 2:
+        v = v+".0"
+    return v
+
+def main():
+    solidity_contracts = list()
+    for root, subdirs, files in os.walk("../../source-code/smart-contract-sanctuary-ethereum"):
+        for file_name in files:
+            if file_name.endswith(".sol"):
+                file_path = os.path.join(root, file_name)
+                solidity_contracts.append(file_path)
+
+    pragmas = dict()
+    for contract in tqdm(solidity_contracts):
+        with open(contract, "r") as f:
+            source_code = f.read()
+            v = extract_version(source_code)
+            if v == None:
+                continue
+            if not v in pragmas:
+                pragmas[v] = 0
+            pragmas[v] += 1
+
+    pragmas = dict(sorted(pragmas.items(), key=lambda item: item[1], reverse=True))
+    for v in pragmas:
+        print(v, "\t", pragmas[v], pragmas[v] / len(solidity_contracts) * 100)
+
+    print()
+    major_versions = dict()
+    for v in pragmas:
+        try:
+            major_version = v.split(".")[0]+"."+v.split(".")[1]
+            if not major_version in major_versions:
+                major_versions[major_version] = 0
+            major_versions[major_version] += pragmas[v]
+        except:
+            pass
+
+    major_versions = dict(sorted(major_versions.items(), key=lambda item: item[1], reverse=True))
+    for v in major_versions:
+        print(v, "\t", major_versions[v], major_versions[v] / len(solidity_contracts) * 100)
+
+    print()
+    unix_timestamp_start_date = int(time.mktime(datetime.datetime(2023, 5, 1, 0, 0).timetuple()))
+    unix_timestamp_end_date = int(time.mktime(datetime.datetime(2023, 6, 30, 0, 0).timetuple()))
+    mongo_connection = pymongo.MongoClient("mongodb://"+MONGO_HOST+":"+str(MONGO_PORT), maxPoolSize=None)
+    collection = mongo_connection["smart_contract_snippets"]["smart_contract_sanctuary_timestamps"]
+    cursor = collection.find({"timestamp":{"$gte":unix_timestamp_start_date,"$lte":unix_timestamp_end_date}})
+    deployments = set()
+    for document in cursor:
+        deployments.add(document["contractAddress"])
+    deployment_versions = dict()
+    for contract in solidity_contracts:
+        address = Web3.toChecksumAddress("0x"+contract.split("/")[-1].split("_")[0])
+        if address.lower() in deployments:
+            with open(contract, "r") as f:
+                source_code = f.read()
+                v = extract_version(source_code)
+                if v != None:
+                    major_version = v.split(".")[0]+"."+v.split(".")[1]
+                    if not major_version in deployment_versions:
+                        deployment_versions[major_version] = 0
+                    deployment_versions[major_version] += 1
+    deployment_versions = dict(sorted(deployment_versions.items(), key=lambda item: item[1], reverse=True))
+    for v in deployment_versions:
+        print(v, "\t", deployment_versions[v], deployment_versions[v] / len(deployments) * 100)
+
+if __name__ == '__main__':
+    main()