Extract and reconstruct cert ids from filenames.

Before this, the regular cert_id regexes were used to extract
the cert_id from the report filename. However, the filenames often
do not use the same cert_id format, but contain all of the information
necessary to reconstruct the cert_id, but with different order for example.

This commit along with those before it introduce a new set of regular
expressions that better match the ones in the filenames. To extract
the correctly formatted canonical cert_id, the regexes are used to
obtain the parts of the cert_id (using named groups in regexes) and
those are then reconstructed into a canonical version of the cert_id
via one of the scheme-dependent functions.

2 files changed, 31 insertions, 25 deletions

diff --git a/src/sec_certs/sample/cc.py b/src/sec_certs/sample/cc.py
index 6ee2bd1b..8948bb76 100644
--- a/src/sec_certs/sample/cc.py
+++ b/src/sec_certs/sample/cc.py
@@ -19,7 +19,7 @@ import sec_certs.utils.pdf
 import sec_certs.utils.sanitization
 from sec_certs import constants
 from sec_certs.cert_rules import SARS_IMPLIED_FROM_EAL, cc_rules, rules, security_level_csv_scan
-from sec_certs.sample.cc_certificate_id import canonicalize
+from sec_certs.sample.cc_certificate_id import canonicalize, schemes
 from sec_certs.sample.certificate import Certificate, References, logger
 from sec_certs.sample.certificate import Heuristics as BaseHeuristics
 from sec_certs.sample.certificate import PdfData as BasePdfData
@@ -337,13 +337,17 @@ class CCCertificate(
             """
             if not self.report_filename:
                 return {}
-            scheme_rules = rules["cc_cert_id"][scheme]
+            scheme_filename_rules = rules["cc_filename_cert_id"][scheme]
+            scheme_meta = schemes[scheme]
             matches: Counter = Counter()
-            for rule in scheme_rules:
+            for rule in scheme_filename_rules:
                 match = re.search(rule, self.report_filename)
                 if match:
-                    cert_id = normalize_match_string(match.group())
-                    matches[cert_id] += 1
+                    try:
+                        cert_id = scheme_meta(match.groupdict())
+                        matches[cert_id] += 1
+                    except Exception:
+                        continue
             if not matches:
                 return {}
             total = max(matches.values())
diff --git a/src/sec_certs/sample/cc_certificate_id.py b/src/sec_certs/sample/cc_certificate_id.py
index ee715082..1fb7b046 100644
--- a/src/sec_certs/sample/cc_certificate_id.py
+++ b/src/sec_certs/sample/cc_certificate_id.py
@@ -182,6 +182,28 @@ def IT(meta) -> str:
     return cert_id
 
 
+# We have rules for some schemes to make canonical cert_ids.
+schemes = {
+    "FR": FR,
+    "DE": DE,
+    "US": US,
+    "MY": MY,
+    "ES": ES,
+    "IN": IN,
+    "SE": SE,
+    "UK": UK,
+    "CA": CA,
+    "JP": JP,
+    "NO": NO,
+    "NL": NL,
+    "AU": AU,
+    "KR": KR,
+    "TR": TR,
+    "SG": SG,
+    "IT": IT,
+}
+
+
 @dataclass(frozen=True)
 class CertificateId:
     """
@@ -210,26 +232,6 @@ class CertificateId:
         """
         The canonical version of this certificate id.
         """
-        # We have rules for some schemes to make canonical cert_ids.
-        schemes = {
-            "FR": FR,
-            "DE": DE,
-            "US": US,
-            "MY": MY,
-            "ES": ES,
-            "IN": IN,
-            "SE": SE,
-            "UK": UK,
-            "CA": CA,
-            "JP": JP,
-            "NO": NO,
-            "NL": NL,
-            "AU": AU,
-            "KR": KR,
-            "TR": TR,
-            "SG": SG,
-            "IT": IT,
-        }
         clean = self.clean
 
         if self.scheme in schemes: