simple_text_extract 3.0.1 → 3.0.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4d97ba86e11c6ad6e14e3e2af600eddc00378747a693a51be38a15da7c73b9b5
-  data.tar.gz: 401dd3aee14d95cc750397bc7b9e6393dbb49eb45d9369d39096a5a6046cb0d8
+  metadata.gz: ddd2289191545f08129aef5ff67b85fe9f2a9c1d1fbe09a377d13afb9203031f
+  data.tar.gz: 99b7a2cb12381cd636115ea350ec4c4c4875e1260ddb7770355a9c6ab0ccf962
 SHA512:
-  metadata.gz: 76156def8c686dca8dba70e7ac36fb58a65757610c7cdb9bfd73c9a13c161d64db1730a21bcf5746a75398cfea2bc1a47e52acfd2e038c414e1a102df05b7854
-  data.tar.gz: f327f06da4692531b051e291f20529889e123c508f15643ade8486819b2ac91ab04bde79b1ed5b6091d3bf8e28c3895fcdbf3cf1148367c16fb2743190aab159
+  metadata.gz: e91131e4458cacd400b751c22dd5d5cf8524ffc4724b4d499fbc522932eca4bde81a507b64ff818382ee620843b9d960921f0f72f970dbf20b09d5315e33515b
+  data.tar.gz: ee5c900bd6ac6cf3f12139976f27aab73dab455d8c73f1fe7e3082291b66a95d31efcaa749225f3f6258a9445539299703e0dde52a0009c8f6a5bd6f75047674

data/-

@@ -0,0 +1 @@
+Test

data/CHANGELOG.md

@@ -1,3 +1,11 @@
+## 3.0.3 (2023-04-27)
+
+- Add support for extracting text from tables in DOCX files.
+
+## 3.0.2 (2023-04-17)
+
+- Coerces filename in `SimpleTextExtract.supports?(filename:)` to string.
+
 ## 3.0.1 (2023-04-17)
 
 - Fixes printing of Roo::Excelx::Cell::Empty for empty rows

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    simple_text_extract (3.0.1)
+    simple_text_extract (3.0.3)
       roo (~> 2.10.0)
       rubyzip (~> 2.3.2)
       spreadsheet (~> 1.3.0)

data/lib/simple_text_extract/extract.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-class SimpleTextExtract::Extract
+class SimpleTextExtract::Extract # rubocop:disable Metrics/ClassLength
   def self.formatter(path)
     case path
     when /.zip$/i
@@ -133,6 +133,13 @@ class SimpleTextExtract::Extract
         doc.xpath("//w:document//w:body/w:p").each do |node|
           result << node.text
         end
+
+        doc.xpath("//w:document//w:body//w:tbl").each do |node|
+          node.xpath(".//w:tr").each do |row|
+            text = row.xpath("w:tc").map(&:text)
+            result << text.join(", ")
+          end
+        end
       end
 
       result.join("\n")

data/lib/simple_text_extract/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module SimpleTextExtract
-  VERSION = "3.0.1"
+  VERSION = "3.0.3"
 end

data/lib/simple_text_extract.rb

@@ -13,7 +13,7 @@ module SimpleTextExtract
   end
 
   def self.supports?(filename: nil)
-    SUPPORTED_FILETYPES.include?(filename.split(".").last)
+    SUPPORTED_FILETYPES.include?(filename.to_s.split(".").last)
   end
 
   def self.missing_dependency?(command)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: simple_text_extract
 version: !ruby/object:Gem::Version
-  version: 3.0.1
+  version: 3.0.3
 platform: ruby
 authors:
 - Nick Weiland
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2023-04-17 00:00:00.000000000 Z
+date: 2023-04-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: roo
@@ -59,6 +59,7 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- "-"
 - ".github/workflows/build.yml"
 - ".gitignore"
 - ".rubocop.yml"