:Qg d dl mZ d dlZd dlmZ d dlmZ d dlmZm Z m
Z
m Z ej e eed g d Z
d Zd
Zd Zy) )partialN)chunk_elements)chunk_by_title)ElementMetadata
NarrativeTextTextTitle)combine_text_under_n_chars)paramsc | j S )N)param)requests h/var/www/html/answerous/venv/lib/python3.12/site-packages/test_unstructured/chunking/test_html_output.py chunking_fnr
s == c , d}d}d}dj |||g }t dt | t dt | t d t | g} | | }t | d
k( sJ |d j
j |k( sJ y ) N%
Header
z@z Header text_as_htmltextmetadatazDate: October 30, 2023z Form field name Example value r )joinr r r lenr r )r
metadata_1
metadata_2
metadata_3combined_metadataelementschunkss r Htest_combining_html_metadata_when_multiple_elements_in_composite_elementr$ s 8JSJ *j*!EF 8o:&NO *_R\5]^ 0?Xb;c
H
"F v;! !9 * *.? ?? ?r c d}d}d}t dt | t dt |d t d
t |d g} | |d }t | d
k( sJ |d j dk( sJ |d j d
k( sJ |d j
j |dz |z k( sJ |d j
j |k( sJ y)aH
Ground truth
r r First1)r parent_idSecond max_characters r r r N)r r r r r r r )r r r r r" r# s r Ftest_combining_html_metadata_with_nested_relationship_between_elementsr/ & s $ 2J9J:J
"J GH?
VY#Z
OWZ$[
H !
4F v;! !9>>W $$ $ !9>>X %% % !9 * *j3.>.K KK K !9 * *j 88 8r c 4 d}t dt | g} | |d }t | dk( sJ |d j d k( sJ |d
j d k( sJ |d j j
dk( sJ |d
j j
dk( sJ y )
z2Mimic behaviour of elements with non-html metadatar r r r r, r. r Hear derN)r r r r r r )r r r" r# s r ;test_html_metadata_exist_in_both_element_when_text_is_splitr4 N s 8J
8o:&NOH !
4F v;! !9>>U "" " !9>>U "" " !9 * *.U UU U !9 * *.U UU Ur ) functoolsr pytestunstructured.chunking.basicr unstructured.chunking.titler unstructured.documents.elementsr r r r fixturer r$ r/ r4 r r r< sV
6 6 W W [\(]^_ `@.%9P Vr