From c7210f2b787fe636cd0aadc5c23c6e292f9f508c Mon Sep 17 00:00:00 2001
From: Roland <roland@learningselfreliance.com>
Date: Mon, 19 Dec 2022 11:27:17 -0700
Subject: [PATCH] Getting title/author from PDF metadata.

---
 test/example.pdf              | Bin 8130 -> 11901 bytes
 wrolpi/files/indexers.py      |  41 +++++++++++++++++++++++++++++-----
 wrolpi/files/test/test_lib.py |  12 +++++-----
 3 files changed, 42 insertions(+), 11 deletions(-)

diff --git a/test/example.pdf b/test/example.pdf
index bc2b9cd2bac8ca8c23e6fb5d3723508e96cf61ec..bbcda4bbd0f32742840804b49ad56ab7a72fe711 100644
GIT binary patch
delta 3848
zcmchaOK;mo5XX^TS}f3WFLt{?A<#i9?Jgz01WJY;BoZKQ4BKhb*MXMS5^ISh$W=xA
zqH}Kr^hE;Xm_vbnjiR5S$DWFImXZP*38m0bN<hN!&*l8)!Nbh&$N&8B;?ILGzV`W?
z$EPvB;xP{fJh5%k_mgPi-Sor3g)b%?i+ILSyS@D5pz<hFY}avJN;Lz1Dy0pu8$GZo
zrMg8aNK|Wp4T4dJAJ~}&dIRi+-bTxR4}STGS&T7DQ(yji$RB*58Oy&vZ7m$N+wEA6
zLzxMVdnW6;9bIp9t&U@Mv>t0Cd*>1SBO(xfZ_EeR$!v-VH5^x^rpq+V?7Mo$W&NIE
zb{tiAy8XUsLV73>UM$iDlmZzB)ljA4Jati}%SV3Tr2f@K7*3u=Lr-wTG)!@L`4h5j
zL@sV3j!DlG-Xt6$a_WtEf(<xuhRD;12c3cN!vKGenba_O{_S@-m`x_DKPqF*bpiOY
z4VmYHMj-OU3r6^e2i0fi4-q-x!h^6p%vMy$g$F|!9@F!2X3gByyD#6Nd88L}PlPdE
zet3BF058em=>i}1n<`aRrf#}grG`SO0s(erVhpE`;&9>x!-ulin*e3RcMG}629fjG
zrO0ag{NxGb%!82_W7RPARkt99>Bq=EP9tw{z2E|0%1*VEzh3>OV1L+BzR^zTNz~(G
z|IuyC&u_oF7~EbDtf6uY-%#32r_mHDhtugKNSbL&@t|hB3AU7pg0myyx}`jxLQk8`
z%}0QlT4TBbPU7KJ^Q`ZJi-q(7w=HEXMARf?KA+d-^;#H@2(v7UP?e}^6;4!>n?QJJ
zHAp@O7YJug`aBuLew2ME(y4bD&O}R55MGBD<y;3zEeo|a2&V)d5LT-%2sp#ohTG<<
z-0fAhZFyoN0gXuaIUHn>Nsxv#UzbICD{(DKoXsCCN6Ei=*WS+G7v(N9e<%6xpHd9j
zo_{{lcd|>TaxZ>qN^$I4_<N*m;r}m`V#t>Ki-RtBQcNj$0p#-h`%udB_e&|@1EhTZ
z1-^>qt*Pvdzm-xPTuJ^tDJA)Pr@W27f|ShvZH~5TyHeQ?e;1`V#4`N*Q_Aq)2jva^
zYe<>=-|(_a*bKQa<HFqBQs#1k1uGG92F_;b1x)*?TwN$;2uAqzg>_A<YxoMwm8HIZ
zV5+)>vjaH58?8}?T|r(s&7e6=#GXIlakiM8`?nl36Opqp6j*LU%frt;*}b}9o7Z=2
rcjZ?%0^94)7*@8rCKrQ>6AU+#`f6RnrrI#q>zynSW169_*Sr4!nexRu

delta 180
zcmewxbI5+fU75*TvdWXoWIH$vEDa4!fM_$DTq+}zq0!`AdE0tZa}!5XV*@uQM;CJ-
zVdiM&WMpCJWMpXI=4#|(>}1DfLr{rAESJ7ZezJ2$YI1gQX|964qlKw~n}w^3xrLLZ
xk(rB|v9U4GcvmMwQzutbS4$TzJ3Fr8lEk8t3jU(hG%j;f3v(`2RabvEE&w1CE{^~J

diff --git a/wrolpi/files/indexers.py b/wrolpi/files/indexers.py
index 5b13695a..865c2b91 100644
--- a/wrolpi/files/indexers.py
+++ b/wrolpi/files/indexers.py
@@ -1,3 +1,4 @@
+import dataclasses
 import pathlib
 import re
 from abc import ABC
@@ -122,6 +123,12 @@ class TextIndexer(Indexer, ABC):
         return words
 
 
+@dataclasses.dataclass
+class PDFMetadata:
+    title: str = None
+    author: str = None
+
+
 @register_indexer('application/pdf')
 class PDFIndexer(Indexer, ABC):
     """Uses PyPDF2 to extract text from a PDF."""
@@ -129,24 +136,47 @@ class PDFIndexer(Indexer, ABC):
     @classmethod
     def create_index(cls, file):
         path: pathlib.Path = file.path.path if hasattr(file.path, 'path') else file.path
-        a = cls.get_title(file)
+
+        file_title = super().get_title(file)
+
+        reader = None
+        if PdfReader is not None:
+            reader = PdfReader(path)
+
+        data = cls.get_pdf_metadata(reader, file)
 
         if path.stat().st_size > FILE_MAX_PDF_SIZE:
             logger.warning(f'PDF too large to fully index: {path}')
-            return a, None, None, None
+            return data.title, data.author, file_title, None
 
         words = ''
         try:
             # PDFs are complex, don't fail to create title index if text extraction fails.
-            words = '\n'.join(truncate_generator_bytes(cls.get_words(path), FILE_MAX_TEXT_SIZE))
+            words = '\n'.join(truncate_generator_bytes(cls.get_words(reader, path), FILE_MAX_TEXT_SIZE))
         except Exception as e:
             logger.error(f'Failed to index {path}', exc_info=e)
             if PYTEST:
                 raise
-        return a, None, None, words
+        return data.title, data.author, file_title, words
 
     @classmethod
-    def get_words(cls, path: pathlib.Path) -> Generator[str, None, None]:
+    def get_pdf_metadata(cls, reader: PdfReader, file) -> PDFMetadata:
+        """Extract title/author from the PDF metadata."""
+        path = file.path.path if hasattr(file.path, 'path') else file.path
+
+        data = PDFMetadata()
+
+        if PdfReader is None:
+            logger.error(f'Cannot get title of {path} PyPDF2 is not installed.')
+            data.title = super().get_title(file)
+            return data
+
+        data.title = reader.metadata.title
+        data.author = reader.metadata.author
+        return data
+
+    @classmethod
+    def get_words(cls, reader: PdfReader, path: pathlib.Path) -> Generator[str, None, None]:
         """
         Reads all text facing up in a PDF.
 
@@ -157,7 +187,6 @@ class PDFIndexer(Indexer, ABC):
             yield ''
             return
 
-        reader = PdfReader(path)
         for page in reader.pages:
             # Get all text facing up.
             text = page.extract_text(0)
diff --git a/wrolpi/files/test/test_lib.py b/wrolpi/files/test/test_lib.py
index 77be95aa..ae1ae2c2 100644
--- a/wrolpi/files/test/test_lib.py
+++ b/wrolpi/files/test/test_lib.py
@@ -488,10 +488,12 @@ def test_pdf_indexer(example_pdf):
 
     a_text, b_text, c_text, d_text = indexers.PDFIndexer.create_index(file)
 
-    # The file name.
-    assert a_text == ['example', 'pdf']
-    assert b_text is None
-    assert c_text is None
+    # The title extracted from the PDF.
+    assert a_text == 'WROLPi Test PDF'
+    # The author.
+    assert b_text == 'roland'
+    # The parsed file name.
+    assert c_text == ['example', 'pdf']
     # Both pages of the example PDF are extracted.
     assert d_text == 'Page one\nPage two'
 
@@ -505,7 +507,7 @@ def test_pdf_indexer_max_size(example_pdf):
     a_text, b_text, c_text, d_text = indexers.PDFIndexer.create_index(file)
 
     # The file name is still indexed.
-    assert a_text == ['example', 'pdf']
+    assert a_text == 'WROLPi Test PDF'
     assert d_text is None