- acc mod

.gitignore

@@ -2,9 +2,6 @@
 data/*/cache.db
 data/*/cache.db-wal
 data/*/cache.db-shm
-output/accuracy/*.db
-output/accuracy/*.db-wal
-output/accuracy/*.db-shm
 #output/export.xlsx
 output/~$export.xlsx
 output/domains/resolve_xx.log

README.md

@@ -44,9 +44,7 @@ uv run scan ch -v             # verbose (streams Docker output)
 
 ## Validation
 
-The classifier is validated via bounce probing — see [`src/mail_municipalities/accuracy/README.md`](src/mail_municipalities/accuracy/README.md) for the full protocol. 
-This is a separate, manual process (not part of the pipeline above) that sends probe emails to a stratified sample of municipalities 
-and parses the resulting NDRs to identify the actual backend MTA.
+The classifier is validated via bounce probing — a separate, manual process (not part of the pipeline above) that sends probe emails to a stratified sample of municipalities and parses the resulting NDRs to identify the actual backend MTA. The validation tooling is maintained in a separate repository (available on request).
 
 ## Output files
 

pyproject.toml

@@ -5,22 +5,19 @@ description = "Domain resolution, provider classification, and security analysis
 readme = "README.md"
 requires-python = ">=3.13"
 dependencies = [
-    "aiosmtplib>=3.0",
     "aiosqlite>=0.20.0",
     "dnspython>=2.8.0",
     "playwright>=1.50",
     "email-validator>=2.3.0",
     "httpx[http2]>=0.28.1",
     "loguru>=0.7.3",
     "pydantic>=2.12.5",
-    "pydantic-settings>=2.0",
     "python-slugify>=8.0.4",
     "stamina>=25.2.0",
     "tldextract>=5.3.1",
     "typer>=0.24.1",
     "pandas>=2.2",
     "openpyxl>=3.1",
-    "scikit-learn>=1.6",
     "seaborn>=0.13.2",
     "scipy>=1.17.1",
 ]
@@ -47,7 +44,6 @@ resolve = "mail_municipalities.cli:resolve"
 classify = "mail_municipalities.cli:classify"
 analyze = "mail_municipalities.cli:analyze"
 scan = "mail_municipalities.cli:scan"
-accuracy = "mail_municipalities.cli:accuracy"
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]