1+ from __future__ import annotations
2+
3+ from abc import ABC , abstractmethod
4+ from typing import Callable , Any
5+
6+ from pptx .presentation import Presentation
7+ from pptx .slide import Slide
8+ from pptx .shapes .base import BaseShape
9+
10+ from ragbits .document_search .documents .document import DocumentMeta
11+ from ragbits .document_search .documents .element import ElementLocation , TextElement
12+
13+
14+ class BasePptxExtractor (ABC ):
15+ """Base class for all PPTX content extractors."""
16+
17+ def _get_slides (self , presentation : Presentation , slide : Slide | None = None ) -> list [tuple [int , Slide ]]:
18+ """Get slides with their indices."""
19+ slides = [slide ] if slide else list (presentation .slides )
20+ return list (enumerate (slides , start = 1 ))
21+
22+ def _create_text_element (
23+ self ,
24+ element_type : str ,
25+ document_meta : DocumentMeta ,
26+ content : str ,
27+ slide_idx : int ,
28+ shape : BaseShape | None = None ,
29+ coordinates : dict [str , Any ] | None = None
30+ ) -> TextElement :
31+ """Create a TextElement with standardized location."""
32+ if coordinates is None and shape is not None :
33+ coordinates = {
34+ "left" : shape .left ,
35+ "top" : shape .top ,
36+ "width" : shape .width ,
37+ "height" : shape .height
38+ }
39+
40+ location = ElementLocation (
41+ page_number = slide_idx ,
42+ coordinates = coordinates or {}
43+ )
44+
45+ return TextElement (
46+ element_type = element_type ,
47+ document_meta = document_meta ,
48+ location = location ,
49+ content = content
50+ )
51+
52+ def _extract_from_shapes (
53+ self ,
54+ presentation : Presentation ,
55+ document_meta : DocumentMeta ,
56+ slide : Slide | None ,
57+ shape_filter : Callable [[BaseShape ], bool ],
58+ content_extractor : Callable [[BaseShape ], str ],
59+ element_type : str = "text"
60+ ) -> list [TextElement ]:
61+ """Generic method to extract content from shapes based on filter and extractor."""
62+ elements : list [TextElement ] = []
63+
64+ for slide_idx , sld in self ._get_slides (presentation , slide ):
65+ for shape in sld .shapes :
66+ if shape_filter (shape ):
67+ try :
68+ content = content_extractor (shape )
69+ if content .strip ():
70+ element = self ._create_text_element (
71+ element_type = element_type ,
72+ document_meta = document_meta ,
73+ content = content ,
74+ slide_idx = slide_idx ,
75+ shape = shape
76+ )
77+ elements .append (element )
78+ except (AttributeError , TypeError ):
79+ continue
80+
81+ return elements
82+
83+ @abstractmethod
84+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
85+ """Extract content from the presentation or specific slide."""
86+
87+ @abstractmethod
88+ def get_extractor_name (self ) -> str :
89+ """Get the name of this extractor."""
90+
91+
92+ class PptxTextExtractor (BasePptxExtractor ):
93+ """Extracts text content from text frames."""
94+
95+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
96+ """Extract text content from the presentation or a specific slide."""
97+ return self ._extract_from_shapes (
98+ presentation = presentation ,
99+ document_meta = document_meta ,
100+ slide = slide ,
101+ shape_filter = lambda shape : shape .has_text_frame ,
102+ content_extractor = lambda shape : str (shape .text_frame .text ).strip ()
103+ )
104+
105+ def get_extractor_name (self ) -> str :
106+ """Get the name of this extractor."""
107+ return "pptx_text_extractor"
108+
109+
110+ class PptxHyperlinkExtractor (BasePptxExtractor ):
111+ """Extracts hyperlink addresses from shapes."""
112+
113+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
114+ """Extract hyperlink content from the presentation or a specific slide."""
115+ return self ._extract_from_shapes (
116+ presentation = presentation ,
117+ document_meta = document_meta ,
118+ slide = slide ,
119+ shape_filter = lambda shape : hasattr (shape , 'click_action' ) and shape .click_action .hyperlink .address ,
120+ content_extractor = lambda shape : shape .click_action .hyperlink .address ,
121+ element_type = "hyperlink"
122+ )
123+
124+ def get_extractor_name (self ) -> str :
125+ """Get the name of this extractor."""
126+ return "pptx_hyperlink_extractor"
127+
128+
129+ class PptxImageExtractor (BasePptxExtractor ):
130+ """Extracts image information from shapes."""
131+
132+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
133+ """Extract image content from the presentation or a specific slide."""
134+ return self ._extract_from_shapes (
135+ presentation = presentation ,
136+ document_meta = document_meta ,
137+ slide = slide ,
138+ shape_filter = lambda shape : shape .image and shape .image is not None ,
139+ content_extractor = lambda shape : f"Image: { shape .image .filename if hasattr (shape .image , 'filename' ) else 'embedded_image' } ,
140+ element_type = "image"
141+ )
142+
143+ def get_extractor_name (self ) -> str :
144+ """Get the name of this extractor."""
145+ return "pptx_image_extractor"
146+
147+
148+ class PptxShapeExtractor (BasePptxExtractor ):
149+ """Extracts shape information and metadata."""
150+
151+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
152+ """Extract shape metadata from the presentation or a specific slide."""
153+ return self ._extract_from_shapes (
154+ presentation = presentation ,
155+ document_meta = document_meta ,
156+ slide = slide ,
157+ shape_filter = lambda shape : hasattr (shape , 'shape_type' ),
158+ content_extractor = lambda shape : f"Shape: { shape .shape_type } ,
159+ element_type = "shape"
160+ )
161+
162+ def get_extractor_name (self ) -> str :
163+ """Get the name of this extractor."""
164+ return "pptx_shape_extractor"
165+
166+
167+ class PptxMetadataExtractor (BasePptxExtractor ):
168+ """Extracts document metadata."""
169+
170+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
171+ """Extract metadata from the presentation."""
172+ core_properties = presentation .core_properties
173+ properties = [
174+ ("author" , core_properties .author ),
175+ ("title" , core_properties .title ),
176+ ("subject" , core_properties .subject ),
177+ ("keywords" , core_properties .keywords ),
178+ ("category" , core_properties .category ),
179+ ("created" , str (core_properties .created ) if core_properties .created else None ),
180+ ("modified" , str (core_properties .modified ) if core_properties .modified else None ),
181+ ]
182+
183+ elements = []
184+ for prop_name , prop_value in properties :
185+ if prop_value is not None and str (prop_value ).strip ():
186+ element = self ._create_text_element (
187+ element_type = "metadata" ,
188+ document_meta = document_meta ,
189+ content = f"{ prop_name } { prop_value } ,
190+ slide_idx = 0
191+ )
192+ elements .append (element )
193+
194+ return elements
195+
196+ def get_extractor_name (self ) -> str :
197+ """Get the name of this extractor."""
198+ return "pptx_metadata_extractor"
199+
200+
201+ class PptxSpeakerNotesExtractor (BasePptxExtractor ):
202+ """Extracts speaker notes from slides."""
203+
204+ def extract (self , presentation : Presentation , document_meta : DocumentMeta , slide : Slide | None = None ) -> list [TextElement ]:
205+ """Extract speaker notes from the presentation or a specific slide."""
206+ elements : list [TextElement ] = []
207+
208+ for slide_idx , sld in self ._get_slides (presentation , slide ):
209+ if sld .has_notes_slide and sld .notes_slide .notes_text_frame is not None :
210+ notes_slide = sld .notes_slide
211+ notes_text_frame = notes_slide .notes_text_frame
212+ text = notes_text_frame .text .strip () if notes_text_frame is not None else None
213+
214+ if text and notes_text_frame is not None :
215+ coordinates = {
216+ "left" : notes_text_frame .margin_left ,
217+ "right" : notes_text_frame .margin_right ,
218+ "top" : notes_text_frame .margin_top ,
219+ "bottom" : notes_text_frame .margin_bottom
220+ }
221+
222+ element = self ._create_text_element (
223+ element_type = "speaker_notes" ,
224+ document_meta = document_meta ,
225+ content = text ,
226+ slide_idx = slide_idx ,
227+ coordinates = coordinates
228+ )
229+ elements .append (element )
230+
231+ return elements
232+
233+ def get_extractor_name (self ) -> str :
234+ """Get the name of this extractor."""
235+ return "pptx_speaker_notes_extractor"
236+
237+
238+ DEFAULT_EXTRACTORS = [
239+ PptxTextExtractor (),
240+ PptxHyperlinkExtractor (),
241+ PptxImageExtractor (),
242+ PptxShapeExtractor (),
243+ PptxSpeakerNotesExtractor (),
244+ PptxMetadataExtractor (),
245+ ]
0 commit comments