From 3b3e78d901a600bb22943202c6a8981ca04a5e48 Mon Sep 17 00:00:00 2001
From: ArztKlein <52363453+ArztKlein@users.noreply.github.com>
Date: Thu, 17 Nov 2022 15:28:46 +1300
Subject: [PATCH] Before and After methods (#175)

* Added before and after functions

* add tests

* formatting
---
 tests/test_cdx_api.py | 36 +++++++++++++++++++
 waybackpy/cdx_api.py  | 82 +++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 118 insertions(+)

diff --git a/tests/test_cdx_api.py b/tests/test_cdx_api.py
index ba2db5a..9748b9f 100644
--- a/tests/test_cdx_api.py
+++ b/tests/test_cdx_api.py
@@ -176,3 +176,39 @@ def test_near() -> None:
             filters=["statuscode:200"],
         )
         cdx.near(unix_timestamp=1286705410)
+
+
+def test_before() -> None:
+    user_agent = (
+        "Mozilla/5.0 (MacBook Air; M1 Mac OS X 11_4) "
+        "AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.1.1 Safari/604.1"
+    )
+
+    cdx = WaybackMachineCDXServerAPI(
+        url="http://www.google.com/",
+        user_agent=user_agent,
+        filters=["statuscode:200"],
+    )
+    before = cdx.before(wayback_machine_timestamp=20160731235949)
+    assert "20160731233347" in before.timestamp
+    assert "google" in before.urlkey
+    assert before.original.find("google.com") != -1
+    assert before.archive_url.find("google.com") != -1
+
+
+def test_after() -> None:
+    user_agent = (
+        "Mozilla/5.0 (MacBook Air; M1 Mac OS X 11_4) "
+        "AppleWebKit/605.1.15 (KHTML, like Gecko) Version/14.1.1 Safari/604.1"
+    )
+
+    cdx = WaybackMachineCDXServerAPI(
+        url="http://www.google.com/",
+        user_agent=user_agent,
+        filters=["statuscode:200"],
+    )
+    after = cdx.after(wayback_machine_timestamp=20160731235949)
+    assert "20160801000917" in after.timestamp, after.timestamp
+    assert "google" in after.urlkey
+    assert after.original.find("google.com") != -1
+    assert after.archive_url.find("google.com") != -1
diff --git a/waybackpy/cdx_api.py b/waybackpy/cdx_api.py
index 6347ad6..4bd8291 100644
--- a/waybackpy/cdx_api.py
+++ b/waybackpy/cdx_api.py
@@ -191,6 +191,88 @@ class WaybackMachineCDXServerAPI:
 
         payload["url"] = self.url
 
+    def before(
+        self,
+        year: Optional[int] = None,
+        month: Optional[int] = None,
+        day: Optional[int] = None,
+        hour: Optional[int] = None,
+        minute: Optional[int] = None,
+        unix_timestamp: Optional[int] = None,
+        wayback_machine_timestamp: Optional[Union[int, str]] = None,
+    ) -> CDXSnapshot:
+        """
+        Gets the nearest archive before the given datetime.
+        """
+        if unix_timestamp:
+            timestamp = unix_timestamp_to_wayback_timestamp(unix_timestamp)
+        elif wayback_machine_timestamp:
+            timestamp = str(wayback_machine_timestamp)
+        else:
+            now = datetime.utcnow().timetuple()
+            timestamp = wayback_timestamp(
+                year=now.tm_year if year is None else year,
+                month=now.tm_mon if month is None else month,
+                day=now.tm_mday if day is None else day,
+                hour=now.tm_hour if hour is None else hour,
+                minute=now.tm_min if minute is None else minute,
+            )
+        self.closest = timestamp
+        self.sort = "closest"
+        self.limit = 25000
+        for snapshot in self.snapshots():
+            if snapshot.timestamp < timestamp:
+                return snapshot
+
+        # If a snapshot isn't returned, then none were found.
+        raise NoCDXRecordFound(
+            "No records were found before the given date for the query."
+            + "Either there are no archives before the given date,"
+            + " the URL may not have any archived, or the URL may have been"
+            + " recently archived and is still not available on the CDX server."
+        )
+
+    def after(
+        self,
+        year: Optional[int] = None,
+        month: Optional[int] = None,
+        day: Optional[int] = None,
+        hour: Optional[int] = None,
+        minute: Optional[int] = None,
+        unix_timestamp: Optional[int] = None,
+        wayback_machine_timestamp: Optional[Union[int, str]] = None,
+    ) -> CDXSnapshot:
+        """
+        Gets the nearest archive after the given datetime.
+        """
+        if unix_timestamp:
+            timestamp = unix_timestamp_to_wayback_timestamp(unix_timestamp)
+        elif wayback_machine_timestamp:
+            timestamp = str(wayback_machine_timestamp)
+        else:
+            now = datetime.utcnow().timetuple()
+            timestamp = wayback_timestamp(
+                year=now.tm_year if year is None else year,
+                month=now.tm_mon if month is None else month,
+                day=now.tm_mday if day is None else day,
+                hour=now.tm_hour if hour is None else hour,
+                minute=now.tm_min if minute is None else minute,
+            )
+        self.closest = timestamp
+        self.sort = "closest"
+        self.limit = 25000
+        for snapshot in self.snapshots():
+            if snapshot.timestamp > timestamp:
+                return snapshot
+
+        # If a snapshot isn't returned, then none were found.
+        raise NoCDXRecordFound(
+            "No records were found after the given date for the query."
+            + "Either there are no archives after the given date,"
+            + " the URL may not have any archives, or the URL may have been"
+            + " recently archived and is still not available on the CDX server."
+        )
+
     def near(
         self,
         year: Optional[int] = None,