From a8ebb2cd246daf04d7b3ec199de10cb094e2b11c Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Guilhem=20Faur=C3=A9?= <pro@gfaure.eu>
Date: Wed, 7 Jun 2023 16:05:09 +0200
Subject: [PATCH] add option to ignore certain objects

---
 spip2md/config.py          |  1 +
 spip2md/extended_models.py | 14 +++++++++++++-
 spip2md/lib.py             |  3 +++
 3 files changed, 17 insertions(+), 1 deletion(-)

diff --git a/spip2md/config.py b/spip2md/config.py
index 38f1513..c544eeb 100644
--- a/spip2md/config.py
+++ b/spip2md/config.py
@@ -31,6 +31,7 @@ class Configuration:
     unknown_char_replacement: str = "??"  # Replaces unknown characters
     clear_log: bool = True  # Clear log before every run instead of appending to
     clear_output: bool = True  # Remove eventual output dir before running
+    ignore_pattern: list[str] = []  # Ignore objects of which title match
     logfile: str = "log-spip2md.log"  # File where logs will be written, relative to wd
     loglevel: str = "WARNING"  # Minimum criticity of logs written in logfile
     logname: str = "spip2md"  # Labelling of logs
diff --git a/spip2md/extended_models.py b/spip2md/extended_models.py
index 7a34f18..58a5dc6 100644
--- a/spip2md/extended_models.py
+++ b/spip2md/extended_models.py
@@ -2,7 +2,7 @@
 import logging
 from os import listdir, makedirs
 from os.path import basename, isfile, splitext
-from re import Match, Pattern, finditer, match, search
+from re import I, Match, Pattern, finditer, match, search
 from shutil import copyfile
 from typing import Any, Optional
 
@@ -301,6 +301,10 @@ class Document(WritableObject, NormalizedDocument):
         return super().write_all(parentdepth, parentdir, index, total)
 
 
+class IgnoredPatternError(Exception):
+    pass
+
+
 class LangNotFoundError(Exception):
     pass
 
@@ -524,6 +528,8 @@ class RedactionalObject(WritableObject):
                 LOG.debug(err)
             except DontExportDraftError as err:
                 LOG.debug(err)
+            except IgnoredPatternError as err:
+                LOG.debug(err)
         return output
 
     # Write object to output destination
@@ -559,6 +565,12 @@ class RedactionalObject(WritableObject):
     # Apply post-init conversions and cancel the export if self not of the right lang
     def convert(self, forced_lang: str) -> None:
         self.convert_title(forced_lang)
+        for p in CFG.ignore_pattern:
+            m = match(p, self._title, I)
+            if m is not None:
+                raise IgnoredPatternError(
+                    f"{self._title} is matching with ignore pattern {p}, ignoring"
+                )
         self.convert_text(forced_lang)
         self.convert_extra()
         if self.lang != forced_lang:
diff --git a/spip2md/lib.py b/spip2md/lib.py
index 423505b..99f7ebc 100644
--- a/spip2md/lib.py
+++ b/spip2md/lib.py
@@ -9,6 +9,7 @@ from spip2md.config import CFG
 from spip2md.extended_models import (
     DeepDict,
     DontExportDraftError,
+    IgnoredPatternError,
     LangNotFoundError,
     Section,
 )
@@ -54,6 +55,8 @@ as database user {esc(BOLD)}{CFG.db_user}{esc()}
                 ROOTLOG.debug(err)  # Log the message
             except DontExportDraftError as err:  # Will happen in not CFG.export_drafts
                 ROOTLOG.debug(err)  # Log the message
+            except IgnoredPatternError as err:
+                ROOTLOG.debug(err)  # Log the message
             print()  # Break line between level 0 sections in output
             ROOTLOG.debug(f"Finished exporting {lang} root section {i}/{nb} {s._title}")
     return {"sections": buffer}