From cda899732c4311ce300d5623cefe9eead55d6e15 Mon Sep 17 00:00:00 2001
From: =?utf8?q?Roland=20H=C3=A4der?= <roland@mxchange.org>
Date: Thu, 25 May 2023 22:07:18 +0200
Subject: [PATCH] Continued: - try to strip off version numbers from software
 name - remove_version() will output a warning and return 'software'
 unmodified if it   fails to match version number against regex

---
 fba.py             | 49 ++++++++++++++++++++++++++++++++++++++++++++--
 fetch_instances.py |  4 ++--
 2 files changed, 49 insertions(+), 4 deletions(-)

diff --git a/fba.py b/fba.py
index b4714dd..502b9b2 100644
--- a/fba.py
+++ b/fba.py
@@ -88,6 +88,40 @@ get_peers_url = "/api/v1/instance/peers"
 connection = sqlite3.connect("blocks.db")
 cursor = connection.cursor()
 
+# Pattern instance for version numbers
+pattern = re.compile("^(?P<version>v|V{0,1})(\.{0,1})(?P<major>0|[1-9]\d*)\.(?P<minor>0|[1-9]\d*)(\.(?P<patch>0|[1-9]\d*)(?:-(?P<prerelease>(?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*)(?:\.(?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*))*))?(?:\+(?P<buildmetadata>[0-9a-zA-Z-]+(?:\.[0-9a-zA-Z-]+)*))?)?$")
+
+def remove_version(software: str) -> str:
+    # NOISY-DEBUG: print(f"DEBUG: software='{software}' - CALLED!")
+    if not "." in software:
+        print(f"WARNING: software='{software}' does not contain a version number.")
+        raise
+
+    version = None
+    if " " in software:
+        version = software.split(" ")[-1]
+    elif "/" in software:
+        version = software.split("/")[-1]
+    elif "-" in software:
+        version = software.split("-")[-1]
+
+    # Run match()
+    matches = pattern.match(version)
+
+    # NOISY-DEBUG: print(f"DEBUG: version[{type(version)}]='{version}',matches='{matches}'")
+    if type(matches) is not re.Match:
+        print(f"WARNING: version='{version}' does not match regex, leaving software='{software}' untouched.")
+        return software
+
+    # NOISY-DEBUG: print(f"DEBUG: Found valid version number: '{version}', removing it ...")
+    end = len(software) - len(version)
+
+    # NOISY-DEBUG: print(f"DEBUG: end[{type(end)}]={end}")
+    software = software[0:end].strip()
+
+    # NOISY-DEBUG: print(f"DEBUG: software='{software}' - EXIT!")
+    return software
+
 def is_blacklisted(domain: str) -> bool:
     blacklisted = False
     for peer in blacklist:
@@ -151,7 +185,7 @@ def update_nodeinfos(domain: str):
             print("WARNING: Did not update any rows:", domain)
 
     except BaseException as e:
-        print(f"ERROR: failed SQL query: domain='{domain}',sql='{sql}',exception='{e}'")
+        print(f"ERROR: failed SQL query: domain='{domain}',sql='{sql}',exception:'{e}'")
         sys.exit(255)
 
     # NOISY-DEBUG: print("DEBUG: Deleting nodeinfos for domain:", domain)
@@ -421,7 +455,7 @@ def fetch_generator_from_path(domain: str, path: str = "/") -> str:
         # NOISY-DEBUG: print(f"DEBUG: Fetching '{path}' from '{domain}' ...")
         res = reqto.get(f"https://{domain}{path}", headers=headers, timeout=(config["connection_timeout"], config["read_timeout"]))
 
-        # NOISY-DEBUG: print("DEBUG: domain,res.ok,res.status_code:", domain, res.ok, res.status_code)
+        # NOISY-DEBUG: print("DEBUG: domain,res.ok,res.status_code,res.text[]:", domain, res.ok, res.status_code, type(res.text))
         if res.ok and res.status_code < 300 and len(res.text) > 0:
             # NOISY-DEBUG: print("DEBUG: Search for <meta name='generator'>:", domain)
             doc = bs4.BeautifulSoup(res.text, "html.parser")
@@ -441,9 +475,17 @@ def fetch_generator_from_path(domain: str, path: str = "/") -> str:
         update_last_error(domain, e)
         pass
 
+    # NOISY-DEBUG: print(f"DEBUG: software[]={type(software)}")
+    if type(software) is str and software == "":
+        software = None
+    elif type(software) is str and "." in software:
+        # NOISY-DEBUG: print(f"DEBUG: software='{software}' may contain a version number, domain='{domain}', removing it ...")
+        software = remove_version(software)
+
     # NOISY-DEBUG: print(f"DEBUG: software='{software}' - EXIT!")
     return software
 
+
 def determine_software(domain: str) -> str:
     # NOISY-DEBUG: print("DEBUG: Determining software for domain:", domain)
     software = None
@@ -496,6 +538,9 @@ def determine_software(domain: str) -> str:
     if str(software) == "":
         # NOISY-DEBUG: print(f"DEBUG: software for '{domain}' was not detected, trying generator ...")
         software = fetch_generator_from_path(domain)
+    elif len(str(software)) > 0 and "." in software:
+        # NOISY-DEBUG: print(f"DEBUG: software='{software}' may contain a version number, domain='{domain}', removing it ...")
+        software = remove_version(software)
 
     # NOISY-DEBUG: print("DEBUG: Returning domain,software:", domain, software)
     return software
diff --git a/fetch_instances.py b/fetch_instances.py
index 9aebc35..20383cf 100644
--- a/fetch_instances.py
+++ b/fetch_instances.py
@@ -53,8 +53,8 @@ def fetch_instances(domain: str, origin: str, software: str):
 
             fba.connection.commit()
 
-        except Exception as e:
-            print("ERROR:", e, instance)
+        except BaseException as e:
+            print(f"ERROR: instance='{instance}',exception:'{e}'")
             continue
 
 instance = sys.argv[1]
-- 
2.39.5