From 1bea64da068c18993d4f4d852a26cdcb2d7f1ba2 Mon Sep 17 00:00:00 2001
From: Adnan Awan <adnan.awan.aa@gmail.com>
Date: Tue, 30 Jan 2024 12:53:12 +0500
Subject: [PATCH 1/2] [dateparser] Fix parsing very short weekday names

- Update en.yaml file to support two letters days of the week
- https://github.com/scrapinghub/dateparser/issues/1170
---
 dateparser/data/date_translation_data/en.py   | 19 +++++++++++------
 .../date_translation_data/en.yaml             | 21 +++++++++++++++++++
 2 files changed, 34 insertions(+), 6 deletions(-)

diff --git a/dateparser/data/date_translation_data/en.py b/dateparser/data/date_translation_data/en.py
index b70975e1b..74fd0c66f 100644
--- a/dateparser/data/date_translation_data/en.py
+++ b/dateparser/data/date_translation_data/en.py
@@ -51,32 +51,39 @@
     ],
     "monday": [
         "mon",
-        "monday"
+        "monday",
+        "mo"
     ],
     "tuesday": [
         "tue",
         "tuesday",
+        "tu",
         "Tues"
     ],
     "wednesday": [
         "wed",
-        "wednesday"
+        "wednesday",
+        "we"
     ],
     "thursday": [
         "thu",
-        "thursday"
+        "thursday",
+        "th"
     ],
     "friday": [
         "fri",
-        "friday"
+        "friday",
+        "fr"
     ],
     "saturday": [
         "sat",
-        "saturday"
+        "saturday",
+        "sa"
     ],
     "sunday": [
         "sun",
-        "sunday"
+        "sunday",
+        "su"
     ],
     "am": [
         "am"
diff --git a/dateparser_data/supplementary_language_data/date_translation_data/en.yaml b/dateparser_data/supplementary_language_data/date_translation_data/en.yaml
index d103bc577..172961545 100644
--- a/dateparser_data/supplementary_language_data/date_translation_data/en.yaml
+++ b/dateparser_data/supplementary_language_data/date_translation_data/en.yaml
@@ -3,9 +3,30 @@ pertain: ["of"]
 
 sentence_splitter_group : 1
 
+# two letters days of week
+
+monday:
+    - mo
+
 tuesday:
+    - tu
     - Tues
 
+wednesday:
+    - we
+
+thursday:
+    - th
+
+friday:
+    - fr
+
+saturday:
+    - sa
+
+sunday:
+    - su
+
 september:
     - sept
 

From 45bd628b5f15c338742b22aa043a6d200dd4efa1 Mon Sep 17 00:00:00 2001
From: Adnan Awan <adnan.awan.aa@gmail.com>
Date: Tue, 30 Jan 2024 23:24:07 +0500
Subject: [PATCH 2/2] [dateparser] Fix parsing very short weekday names

- Develop method to remove_multiple_occurrences of the day(s) of the week
- https://github.com/scrapinghub/dateparser/issues/1170
---
 dateparser/languages/locale.py | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/dateparser/languages/locale.py b/dateparser/languages/locale.py
index bd0df2c66..7ef489974 100644
--- a/dateparser/languages/locale.py
+++ b/dateparser/languages/locale.py
@@ -109,6 +109,31 @@ def clean_dictionary(dictionary, threshold=2):
             del dictionary[del_key]
         return dictionary
 
+    @property
+    def weekdays(self):
+        weekdays = [
+            "monday",
+            "tuesday",
+            "wednesday",
+            "thursday",
+            "friday",
+            "saturday",
+            "sunday",
+        ]
+        return weekdays
+
+    def remove_multiple_occurrences(self, date_str_tokens: list):
+        # first occurrence of day of the week will be considered
+        # followings occurrence(s) will be skipped and removed from the token list.
+        weekdays_counter = 0
+        for i, token in enumerate(date_str_tokens):
+            if token in self.weekdays:
+                weekdays_counter += 1
+
+                if weekdays_counter > 1:
+                    date_str_tokens.pop(i)
+                    continue
+
     def translate(self, date_string, keep_formatting=False, settings=None):
         """
         Translate the date string to its English equivalent.
@@ -145,6 +170,7 @@ def translate(self, date_string, keep_formatting=False, settings=None):
         if "in" in date_string_tokens:
             date_string_tokens = self._clear_future_words(date_string_tokens)
 
+        self.remove_multiple_occurrences(date_string_tokens)
         return self._join(
             list(filter(bool, date_string_tokens)),
             separator="" if keep_formatting else " ",