9047196931

Committed 11 May 2024 11:44PM UTC coverage: 73.684% (-0.7%) from 74.359%

Build # 9047196931

Build Type

Pull #49

github

Committed by

lfoppiano

Commit Message

Harmonise how the width is collected, update tests

Pull Request Pull Request #49: Fix scaling and display high PDF in high resolution

Run Details

2 of 2 new or added lines in 1 file covered. (100.0%)

6 existing lines in 2 files now uncovered.

56 of 76 relevant lines covered (73.68%)

8.84 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

73.68

/streamlit_pdf_viewer/__init__.py

import base64
import os
from pathlib import Path
from typing import Union, List, Optional

import streamlit.components.v1 as components
import json

_RELEASE = True
RENDERING_EMBED = "legacy_embed"
RENDERING_IFRAME = "legacy_iframe"
RENDERING_UNWRAP = "unwrap"

if not _RELEASE:
    _component_func = components.declare_component(
        "streamlit_pdf_viewer",
        url="http://localhost:3001",
    )
else:
    parent_dir = os.path.dirname(os.path.abspath(__file__))
    build_dir = os.path.join(parent_dir, "frontend/dist")
    _component_func = components.declare_component(
        "streamlit_pdf_viewer",
        path=build_dir
    )


def pdf_viewer(input: Union[str, Path, bytes], 
               width: int = None,
               height: int = None, 
               key=None,
               annotations: list = (),
               pages_vertical_spacing: int = 2,
               annotation_outline_size: int = 1,
               rendering: str = RENDERING_UNWRAP,
               pages_to_render: List[int] = ()
               ):
    """
    pdf_viewer function to display a PDF file in a Streamlit app.

    :param input: The source of the PDF file. Accepts a file path, URL, or binary data.
    :param width: Width of the PDF viewer in pixels. Defaults to 700 pixels.
    :param height: Height of the PDF viewer in pixels. If not provided, the viewer show the whole content.
    :param key: An optional key that uniquely identifies this component. Used to preserve state in Streamlit apps.
    :param annotations: A list of annotations to be overlaid on the PDF. Each annotation should be a dictionary.
    :param pages_vertical_spacing: The vertical space (in pixels) between each page of the PDF. Defaults to 2 pixels.
    :param annotation_outline_size: Size of the outline around each annotation in pixels. Defaults to 1 pixel.
    :param rendering: Type of rendering. The default is "unwrap", which unwrap the PDF. Other values are
    :param pages_to_render: Optional list of page numbers to render. If None, all pages are rendered. This allows for selective rendering of pages in the PDF.
    "legacy_iframe" and "legacy_embed" which uses the legacy approach for showing PDF document with streamlit.
    These methods enable the default pdf viewer of Firefox/Chrome/Edge that contains additional features we are still
    working to implement for the "unwrap" method.

    The function reads the PDF file (from a file path, URL, or binary data), encodes it in base64,
    and uses a Streamlit component to render it in the app. It supports optional annotations and adjustable margins.

    Returns the value of the selected component (if any).
    """

    # Validate width and height parameters
    if width is not None and not isinstance(width, int):
        raise TypeError("Width must be an integer")
    if height is not None and not isinstance(height, int):
        raise TypeError("Height must be an integer or None")
    if not all(isinstance(page, int) for page in pages_to_render):
        raise TypeError("pages_to_render must be a list of integers")

    if type(input) is not bytes:
        with open(input, 'rb') as fo:
            binary = fo.read()
    else:
        binary = input
        
    if rendering == RENDERING_IFRAME or rendering == RENDERING_EMBED:
        if height is None: 
            height = "100%"

    base64_pdf = base64.b64encode(binary).decode('utf-8')
    component_value = _component_func(
        binary=base64_pdf,
        width=width,
        height=height,
        key=key,
        default=0,
        annotations=annotations,
        pages_vertical_spacing=pages_vertical_spacing,
        annotation_outline_size=annotation_outline_size,
        rendering=rendering,
        pages_to_render=pages_to_render
    )
    return component_value


if not _RELEASE:
    with open("resources/test.pdf", 'rb') as fo:
        binary = fo.read()

    with open("resources/annotations.json", 'rb') as fo:
        annotations = json.loads(fo.read())

    viewer = pdf_viewer(
        binary,
        # height=700,
        width=800,
        annotations=annotations
    )

1	import base64	12✔
2	import os	12✔
3	from pathlib import Path	12✔
4	from typing import Union, List, Optional	12✔
5
6	import streamlit.components.v1 as components	12✔
7	import json	12✔
8
9	_RELEASE = True	12✔
10	RENDERING_EMBED = "legacy_embed"	12✔
11	RENDERING_IFRAME = "legacy_iframe"	12✔
12	RENDERING_UNWRAP = "unwrap"	12✔
13
14	if not _RELEASE:	12✔
15	_component_func = components.declare_component(	×
16	"streamlit_pdf_viewer",
17	url="http://localhost:3001",
18	)
19	else:
20	parent_dir = os.path.dirname(os.path.abspath(__file__))	12✔
21	build_dir = os.path.join(parent_dir, "frontend/dist")	12✔
22	_component_func = components.declare_component(	12✔
23	"streamlit_pdf_viewer",
24	path=build_dir
25	)
26
27
28	def pdf_viewer(input: Union[str, Path, bytes],	12✔
29	width: int = None,
30	height: int = None,
31	key=None,
32	annotations: list = (),
33	pages_vertical_spacing: int = 2,
34	annotation_outline_size: int = 1,
35	rendering: str = RENDERING_UNWRAP,
36	pages_to_render: List[int] = ()
37	):
38	"""
39	pdf_viewer function to display a PDF file in a Streamlit app.
40
41	:param input: The source of the PDF file. Accepts a file path, URL, or binary data.
42	:param width: Width of the PDF viewer in pixels. Defaults to 700 pixels.
43	:param height: Height of the PDF viewer in pixels. If not provided, the viewer show the whole content.
44	:param key: An optional key that uniquely identifies this component. Used to preserve state in Streamlit apps.
45	:param annotations: A list of annotations to be overlaid on the PDF. Each annotation should be a dictionary.
46	:param pages_vertical_spacing: The vertical space (in pixels) between each page of the PDF. Defaults to 2 pixels.
47	:param annotation_outline_size: Size of the outline around each annotation in pixels. Defaults to 1 pixel.
48	:param rendering: Type of rendering. The default is "unwrap", which unwrap the PDF. Other values are
49	:param pages_to_render: Optional list of page numbers to render. If None, all pages are rendered. This allows for selective rendering of pages in the PDF.
50	"legacy_iframe" and "legacy_embed" which uses the legacy approach for showing PDF document with streamlit.
51	These methods enable the default pdf viewer of Firefox/Chrome/Edge that contains additional features we are still
52	working to implement for the "unwrap" method.
53
54	The function reads the PDF file (from a file path, URL, or binary data), encodes it in base64,
55	and uses a Streamlit component to render it in the app. It supports optional annotations and adjustable margins.
56
57	Returns the value of the selected component (if any).
58	"""
59
60	# Validate width and height parameters
61	if width is not None and not isinstance(width, int):	12✔
62	raise TypeError("Width must be an integer")	×
63	if height is not None and not isinstance(height, int):	12✔
64	raise TypeError("Height must be an integer or None")	×
65	if not all(isinstance(page, int) for page in pages_to_render):	12✔
66	raise TypeError("pages_to_render must be a list of integers")	×
67
68	if type(input) is not bytes:	12✔
69	with open(input, 'rb') as fo:	12✔
70	binary = fo.read()	12✔
71	else:
72	binary = input	×
73
74	if rendering == RENDERING_IFRAME or rendering == RENDERING_EMBED:	12✔
75	if height is None:	12✔
76	height = "100%"	12✔
77
78	base64_pdf = base64.b64encode(binary).decode('utf-8')	12✔
79	component_value = _component_func(	12✔
80	binary=base64_pdf,
81	width=width,
82	height=height,
83	key=key,
84	default=0,
85	annotations=annotations,
86	pages_vertical_spacing=pages_vertical_spacing,
87	annotation_outline_size=annotation_outline_size,
88	rendering=rendering,
89	pages_to_render=pages_to_render
90	)
91	return component_value	12✔
92
93
94	if not _RELEASE:	12✔
UNCOV 95	with open("resources/test.pdf", 'rb') as fo:	×
UNCOV 96	binary = fo.read()	×
97
98	with open("resources/annotations.json", 'rb') as fo:	×
UNCOV 99	annotations = json.loads(fo.read())	×
100
101	viewer = pdf_viewer(	×
102	binary,
103	# height=700,
104	width=800,
105	annotations=annotations
106	)

lfoppiano / streamlit-pdf-viewer / 9047196931

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous