From a02e7dcbce9013a1ba6b376dfe0acbb58929c76d Mon Sep 17 00:00:00 2001
From: "Gerber, Mike" <mike.gerber@sbb.spk-berlin.de>
Date: Wed, 10 Jun 2020 18:29:11 +0200
Subject: [PATCH] =?UTF-8?q?=F0=9F=9A=A7=20dinglehopper:=20WIP=20data=20str?=
 =?UTF-8?q?ucture=20for=20extracted=20text?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 extracted_text.py | 42 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 extracted_text.py

diff --git a/extracted_text.py b/extracted_text.py
new file mode 100644
index 0000000..d1dc6f0
--- /dev/null
+++ b/extracted_text.py
@@ -0,0 +1,42 @@
+import attr
+
+
+@attr.s(frozen=True)
+class ExtractedText:
+    segments = attr.ib()
+    joiner = attr.ib(type=str)
+    # XXX Use type annotations for attr types when support for Python 3.5 is dropped
+    # XXX Also I think these are not validated?
+
+    @property
+    def text(self):
+        return self.joiner.join(s.text for s in self.segments)
+
+    def segment_id_for_pos(self, pos):
+        i = 0
+        for s in self.segments:
+            if i <= pos < i + len(s.text):
+                return s.id
+            i += len(s.text)
+            if i <= pos < i + len(self.joiner):
+                return None
+            i += len(self.joiner)
+
+
+@attr.s(frozen=True)
+class ExtractedTextSegment:
+    id = attr.ib(type=str)
+    text = attr.ib(type=str)
+
+
+test1 = ExtractedText([
+    ExtractedTextSegment('s0', 'foo'),
+    ExtractedTextSegment('s1', 'bar'),
+    ExtractedTextSegment('s2', 'bazinga')
+], ' ')
+
+
+assert test1.text == 'foo bar bazinga'
+assert test1.segment_id_for_pos(0) == 's0'
+assert test1.segment_id_for_pos(3) == None
+assert test1.segment_id_for_pos(10) == 's2'