Fix object coordinates when mediabox != (0,0,...)

See issue #1181 for details.

CHANGELOG.md

@@ -16,6 +16,7 @@ All notable changes to this project will be documented in this file. The format
 
 ### Fixed
 
+- Fix handling of object coordinates when `mediabox` does not begin at `(0,0)` (h/t @wodny). ([#1181](https://github.com/jsvine/pdfplumber/issues/1181))
 - Fix error on getting `.annots`/`.hyperlinks` from `CroppedPage` (due to missing `.rotation` and `.initial_doctop` attributes) (h/t @Safrone). ([#1171](https://github.com/jsvine/pdfplumber/issues/1171) + [e5737d2](https://github.com/jsvine/pdfplumber/commit/e5737d2))
 - Fix problem where `Page.crop(...)` was not cropping `.annots/.hyperlinks` (h/t @Safrone). [#1171](https://github.com/jsvine/pdfplumber/issues/1171)
 - Fix calculation of coordinates for `.annots` on `CroppedPage`s. ([0bbb340](https://github.com/jsvine/pdfplumber/commit/0bbb340))

README.md

@@ -568,6 +568,7 @@ Many thanks to the following users who've contributed ideas, features, and fixes
 - [Aron Weiler](https://github.com/aronweiler)
 - [Quentin André](https://github.com/QuentinAndre11)
 - [Léo Roux](https://github.com/leorouxx)
+- [@wodny](https://github.com/wodny)
 
 ## Contributing
 

pdfplumber/page.py

@@ -415,11 +415,20 @@ def process_attr(item: Tuple[str, Any]) -> Optional[Tuple[str, Any]]:
 
             attr["dash"] = obj.dashing_style
 
+        # As noted in #1181, `pdfminer.six` adjusts objects'
+        # coordinates relative to the MediaBox:
+        # https://github.com/pdfminer/pdfminer.six/blob/1a8bd2f730295b31d6165e4d95fcb5a03793c978/pdfminer/converter.py#L79-L84
+        mb_x0, mb_top = self.mediabox[:2]
+
         if "y0" in attr:
-            attr["top"] = self.height - attr["y1"]
-            attr["bottom"] = self.height - attr["y0"]
+            attr["top"] = (self.height - attr["y1"]) + mb_top
+            attr["bottom"] = (self.height - attr["y0"]) + mb_top
             attr["doctop"] = self.initial_doctop + attr["top"]
 
+        if "x0" in attr and mb_x0 != 0:
+            attr["x0"] = attr["x0"] + mb_x0
+            attr["x1"] = attr["x1"] + mb_x0
+
         return attr
 
     def iter_layout_objects(

tests/test_issues.py

@@ -310,3 +310,25 @@ def test_issue_1147(self):
             page = pdf.pages[0]
             # Should not error:
             assert page.extract_text()
+
+    def test_issue_1181(self):
+        """
+        Correctly re-calculate coordinates when MediaBox does not start at (0,0)
+        """
+        path = os.path.join(HERE, "pdfs/issue-1181.pdf")
+        with pdfplumber.open(path) as pdf:
+            p0, p1 = pdf.pages
+            assert p0.crop(p0.bbox).extract_table() == [
+                ["FooCol1", "FooCol2", "FooCol3"],
+                ["Foo4", "Foo5", "Foo6"],
+                ["Foo7", "Foo8", "Foo9"],
+                ["Foo10", "Foo11", "Foo12"],
+                ["", "", ""],
+            ]
+            assert p1.crop(p1.bbox).extract_table() == [
+                ["BarCol1", "BarCol2", "BarCol3"],
+                ["Bar4", "Bar5", "Bar6"],
+                ["Bar7", "Bar8", "Bar9"],
+                ["Bar10", "Bar11", "Bar12"],
+                ["", "", ""],
+            ]