pdftotext 0.0.1.0 → 0.0.2.0

raw patch · 9 files changed

+163/−9 lines, 9 filesdep +xml-conduitPVP ok

version bump matches the API change (PVP)

Dependencies added: xml-conduit

API changes (from Hackage documentation)

Files

• CHANGELOG.md +5/−0
• README.md +6/−0
• cbits/poppler.cc +36/−7
• pdftotext.cabal +8/−1
• src/Pdftotext.hs +12/−0
• src/Pdftotext/Foreign.hs +29/−1
• src/Pdftotext/Internal.hs +56/−0
• test/PdftotextSpec.hs +11/−0
• test/simple.pdf binary

@@ -1,5 +1,10 @@ # Revision history for pdftotext +## 0.0.2.0 -- 2020-06-11++* Added PDF document properties (author, title etc.)+* Added flag `xml-conduit` (parse metadata using `xml-conduit`)+ ## 0.0.1.0 -- 2020-05-10  * First version.

@@ -14,6 +14,12 @@   T.putStrLn $ pdftotext Physical pdf ``` +## Flags++### `xml-conduit`++`pdftotext` can extract properties from PDF document. One of them is metadata which is in form of XML document. If `xml-conduit` flag is set then the metadata is parsed using `xml-conduit` package, otherwise provided as text.+ ## Internals  The library uses poppler via FFI, therefore internally all functions are of type `IO`. However, their non-`IO` variants (using `unsafePerformIO`) _should be_ safe to use. Module `Pdftotext.Internal` exposes all `IO`-typed functions.

@@ -5,6 +5,11 @@  extern "C" { +  std::string* to_stdstring(poppler::ustring ust) {+    std::vector<char> vc = ust.to_utf8();+    return new std::string(vc.begin(), vc.end());+  }+     poppler::document* poppler_document_open_pdf(const char* file) {     poppler::document* doc = poppler::document::load_from_file(file);     return doc;@@ -24,6 +29,34 @@     return doc->pages();   } +  std::string* poppler_document_author(poppler::document* doc) {+    return to_stdstring(doc->get_author());+  }++  std::string* poppler_document_creator(poppler::document* doc) {+    return to_stdstring(doc->get_creator());+  }++  std::string* poppler_document_producer(poppler::document* doc) {+    return to_stdstring(doc->get_producer());+  }++  std::string* poppler_document_subject(poppler::document* doc) {+    return to_stdstring(doc->get_subject());+  }++  std::string* poppler_document_title(poppler::document* doc) {+    return to_stdstring(doc->get_title());+  }++  std::string* poppler_document_keywords(poppler::document* doc) {+    return to_stdstring(doc->get_keywords());+  }++  std::string* poppler_document_metadata(poppler::document* doc) {+    return to_stdstring(doc->metadata());+  }+   poppler::page* poppler_document_open_page(poppler::document* doc, int page) {     return doc->create_page(page);   }@@ -32,19 +65,15 @@     std::vector<char> vc;     switch (layout) {     case 0: { -      vc = page->text(poppler::rectf(), poppler::page::text_layout_enum::raw_order_layout).to_utf8();-      break;+      return to_stdstring(page->text(poppler::rectf(), poppler::page::text_layout_enum::raw_order_layout));     }     case 1: {-      vc = page->text(poppler::rectf(), poppler::page::text_layout_enum::physical_layout).to_utf8();-      break;+      return to_stdstring(page->text(poppler::rectf(), poppler::page::text_layout_enum::physical_layout));     }     default: {-      vc = page->text(poppler::rectf(), poppler::page::text_layout_enum::non_raw_non_physical_layout).to_utf8();-      break;+      return to_stdstring(page->text(poppler::rectf(), poppler::page::text_layout_enum::non_raw_non_physical_layout));     }     }-    return new std::string(vc.begin(), vc.end());   }    void poppler_page_delete(poppler::page* page) {

@@ -1,7 +1,7 @@ cabal-version:       >=1.10  name:                pdftotext-version:             0.0.1.0+version:             0.0.2.0 synopsis:            Extracts text from PDF using poppler description:         The @pdftotext@ package provides functions for extraction of plain text from PDF documents. It uses C++ library [Poppler](https://poppler.freedesktop.org/), which is required to be installed in the system. Output of Haskell @pdftotext@ library is identical to output of Poppler's tool @pdftotext@. homepage:            https://sr.ht/~geyaeb/haskell-pdftotext/@@ -19,6 +19,10 @@   type:                mercurial   location:            https://hg.sr.ht/~geyaeb/haskell-pdftotext +flag xml-conduit+  description:         Parse metadata of PDF document properties using xml-conduit+  default:             False+   library   exposed-modules:     Pdftotext                      , Pdftotext.Foreign@@ -34,6 +38,9 @@                      , cbits/stdstring.cc   extra-libraries:     stdc++   pkgconfig-depends:   poppler-cpp+  if flag(xml-conduit)+     build-depends:    xml-conduit == 1.8.*+     cpp-options:      -DXMLC  test-suite pdftotext-test   default-language:    Haskell2010

@@ -20,6 +20,10 @@ >   Just pdf <- openFile "path/to/file.pdf" >   T.putStrLn $ pdftotext Physical pdf +=== Flags++* @xml-conduit@ – 'metadata' of PDF document properties is parsed as XML, otherwise remains as text+ -} {- ORMOLU_ENABLE -} module Pdftotext@@ -27,6 +31,7 @@     Document,     Layout (..),     Page,+    Properties (..),      -- * Loading PDF's     openByteString,@@ -37,6 +42,7 @@     pages,     pagesTotal,     pdftotext,+    properties,      -- * Page functions     pageNumber,@@ -66,6 +72,12 @@ -- | Return number of pages contained in document. pagesTotal :: Document -> Int pagesTotal = unsafePerformIO . pagesTotalIO++-- | Extract properties from the document.+--+-- @since 0.0.2.0+properties :: Document -> Properties+properties = unsafePerformIO . propertiesIO  -- | Extract text from PDF document with given 'Layout'. pdftotext :: Layout -> Document -> Text

@@ -24,9 +24,16 @@     -- * FFI     ffiOpenPdf,     ffiOpenData,+    ffiDocumentAuthor,+    ffiDocumentCreator,     ffiDocumentDelete,-    ffiDocumentPages,+    ffiDocumentKeywords,+    ffiDocumentMetadata,     ffiDocumentOpenPage,+    ffiDocumentPages,+    ffiDocumentProducer,+    ffiDocumentSubject,+    ffiDocumentTitle,     ffiPageDelete,     ffiPageText,     ffiStringLength,@@ -59,6 +66,27 @@  foreign import ccall unsafe "poppler_document_pages"   ffiDocumentPages :: Ptr Poppler_Document -> IO CInt++foreign import ccall unsafe "poppler_document_author"+  ffiDocumentAuthor :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_creator"+  ffiDocumentCreator :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_producer"+  ffiDocumentProducer :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_subject"+  ffiDocumentSubject :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_title"+  ffiDocumentTitle :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_keywords"+  ffiDocumentKeywords :: Ptr Poppler_Document -> IO StdString++foreign import ccall unsafe "poppler_document_metadata"+  ffiDocumentMetadata :: Ptr Poppler_Document -> IO StdString  foreign import ccall unsafe "poppler_document_open_page"   ffiDocumentOpenPage :: Ptr Poppler_Document -> CInt -> IO (Ptr Poppler_Page)

@@ -1,4 +1,5 @@ {-# LANGUAGE BlockArguments #-}+{-# LANGUAGE CPP #-}  {- ORMOLU_DISABLE -} {-|@@ -18,6 +19,7 @@     Document (..),     Layout (..),     Page (..),+    Properties (..),      -- * Loading PDF's     openByteStringIO,@@ -28,6 +30,7 @@     pagesIO,     pagesTotalIO,     pdftotextIO,+    propertiesIO,      -- * Page functions     pageTextIO,@@ -41,8 +44,33 @@ import Foreign.C (withCString) import Pdftotext.Foreign +#ifdef XMLC+import qualified Text.XML as X+import qualified Data.Text.Lazy as TL+#endif+ newtype Document = Document (ForeignPtr Poppler_Document) +-- | Document properties.+--+-- If flag @xml-conduit@ is set, 'metadata' is of type @Maybe Text.XML.Document@.+--+-- @since 0.0.2.0+data Properties = Properties+  { author :: Maybe T.Text,+    creator :: Maybe T.Text,+    keywords :: Maybe T.Text,+#ifdef XMLC+    metadata :: Maybe X.Document,+#else+    metadata :: Maybe T.Text,+#endif+    producer :: Maybe T.Text,+    subject :: Maybe T.Text,+    title :: Maybe T.Text+  }+  deriving (Show)+ data Page = Page   { -- | Number of this page in original document.     pageNumber :: Int,@@ -118,6 +146,34 @@         Raw -> 0         Physical -> 1         None -> 2++-- | Extract properties from the document.+-- @since 0.0.2.0+propertiesIO :: Document -> IO Properties+propertiesIO (Document docptr) = withForeignPtr docptr \doc -> do+  a <- asText $ ffiDocumentAuthor doc+  c <- asText $ ffiDocumentCreator doc+  k <- asText $ ffiDocumentKeywords doc+  m <- asText $ ffiDocumentMetadata doc+  p <- asText $ ffiDocumentProducer doc+  s <- asText $ ffiDocumentSubject doc+  t <- asText $ ffiDocumentTitle doc++#ifdef XMLC+  return $ Properties (f a) (f c) (f k) (xml m) (f p) (f s) (f t)+  where+    xml x =+      if T.null x+      then Nothing+      else either (const Nothing) Just $ X.parseText X.def (TL.fromStrict x)+#else+  return $ Properties (f a) (f c) (f k) (f m) (f p) (f s) (f t)+  where+#endif+    f x =+      if T.null x+      then Nothing+      else Just x  -- | Extract text from PDF document with given 'Layout'. pdftotextIO :: Layout -> Document -> IO T.Text

@@ -1,5 +1,6 @@ {-# LANGUAGE BlockArguments #-} {-# LANGUAGE OverloadedStrings #-}+{-# LANGUAGE RecordWildCards #-}  module PdftotextSpec (spec) where @@ -30,3 +31,13 @@         pagesTotal doc `shouldBe` 4       it "should contain correct number of pages (`pages`)" \(Just doc) ->         length (pages doc) `shouldBe` 4++    describe "PDF properties" do+      it "should contain all fields" \(Just doc) -> do+        let Properties {..} = properties doc+        author `shouldBe` Just "G. Eyaeb"+        title `shouldBe` Just "Simple document for testing"+        creator `shouldBe` Just "pdflatex"+        producer `shouldBe` Just "LaTeX with hyperref"+        keywords `shouldBe` Just "haskell,pdf"+        subject `shouldBe` Just "Testing"

binary file changed (25660 → 25932 bytes)