NgC UdZddlmZddlZddlmZddlmZmZm Z m Z ddl Z ddl ZddlZddlmZmZddlmZddlmZdd lmZdd lmZmZmZmZmZm Z m!Z!dd l"m#Z#dd l$m%Z%m&Z&dd l'm(Z(m)Z)m*Z*m+Z+ddl,m-Z-dZ.de/d<dZ0de/d<e%e#j1e d8ddddddd9d'Z2Gd(d)Z3Gd*d+Z4Gd,d-Z5Gd.d/Z6d:d2Z7d;d7Z8dS) open(filename, "rb"). find_subtable Detect "subtables" on each worksheet and partition each of those as a separate `Table` element. When `False`, each worksheet is partitioned as a single `Table` element. A subtable is a contiguous block of cells with more than two cells in each row. infer_table_structure If True, any Table elements that are extracted will also have a metadata field named "text_as_html" where the table's text content is rendered into an html string. I.e., rows and cells are preserved. Whether True or False, the "text" field is always present in any Table element and is the text content of the table (no structure). include_header Determines whether or not header info is included in text and medatada.text_as_html ) file_pathr"r#r$r%)startF)indexheaderna_repN) text_as_html page_name page_numberr' last_modified)textmetadatar:)_XlsxPartitionerOptions enumeratesheetsitemsr#r from_html_textto_htmlr$rhtmlmetadata_file_pathr9r detection_originappendrr:_ConnectedComponentsfrom_worksheet_df_SubtableParsersubtable#iter_leading_single_cell_rows_texts_create_elementr _get_metadatar; core_tabler%r6$iter_trailing_single_cell_rows_texts)r'r"r#r$r%r&r,optselementsr8 sheet_namesheet html_tabler; componentsubtable_parsercontentelementrNs W/var/www/html/ai-engine/env/lib/python3.11/site-packages/unstructured/partition/xlsx.pypartition_xlsxrZ)sB # #%3    D!H,5 #7---0-0-( (j%!- -"1 E$2Eb QQJ'0EOZ__4$'0"0 H)9H % OOEzJJJ K K K K2CCEJJ - - "1)2D"E"E /RRTT--G-c'll;;G'4Zd'S'SG$OOG,,,,-7 )!*!9"**t?R[]*^^""J$999G'4Zd'S'SG$+/+EO 4$1OOG,,, /SSUU--G-c'll;;G'4Zd'S'SG$OOG,,,,-3 -< OceZdZdZdd Zedd ZeddZeddZeddZ eddZ eddZ eddZ dS)r=zVEncapsulates partitioning option validation, computation, and application of defaults.r0r(r"r)r#r*r$r%cL||_||_||_||_||_dSN) _file_path_file_find_subtable_include_header_infer_table_structure)selfr0r"r#r$r%s rY__init__z _XlsxPartitionerOptions.__init__s0$ +-&;###r[r-c|jS)zTrue when partitioner should detect and emit separate `Table` elements for subtables. A subtable is (roughly) a contiguous rectangle of populated cells bounded by empty rows. )rards rYr#z%_XlsxPartitionerOptions.find_subtables ""r[ int | Nonec|jrdndS)zNThe index of the row Pandas should treat as column-headings. Either 0 or None.rNrbrgs rYheader_row_idxz&_XlsxPartitionerOptions.header_row_idxs(2qqd2r[c|jS)z6True when column headers should be included in tables.rjrgs rYr$z&_XlsxPartitionerOptions.include_headers ##r[c|jS)zGTrue when partitioner should compute and apply `text_as_html` metadata.)rcrgs rYr%z-_XlsxPartitionerOptions.infer_table_structures **r[c<|jrt|jndS)zHThe best last-modified date available, None if no sources are available.N)r_rrgs rYr9z%_XlsxPartitionerOptions.last_modifieds!;?/S%do666tSr[ str | Nonec|jS)zHThe best available file-path for this document or `None` if unavailable.)r_rgs rYrDz*_XlsxPartitionerOptions.metadata_file_paths r[dict[str, pd.DataFrame]cT|jx}rtj|d|jS|jx}rlt |t r;|dtj | }tj|d|jStd)zFThe spreadsheet worksheets, each as a data-frame mapped by sheet-name.N)rRr4rz7Either 'filename' or 'file' argument must be specified.) r_pd read_excelrkr` isinstancerseekioBytesIOread ValueError)rdr0fs rYr?z_XlsxPartitionerOptions.sheetss '9 Y=tDDWXXX X ?1 Q!122 )q Jqvvxx((=tD > >\ > > >r[rcZeZdZdZddZddZedd Zedd Z dd Z dS)rGzThe collection of connected-components for a single worksheet. "Connected-components" refers to the graph algorithm we use to detect contiguous groups of non-empty cells in an excel sheet. worksheet_dfrc||_dSr^) _worksheet_df)rdrs rYrez_ConnectedComponents.__init__ s)r[r-Iterator[_ConnectedComponent]c*t|jSr^)iter_connected_componentsrgs rY__iter__z_ConnectedComponents.__iter__ sD.///r[r c||S)zKConstruct from a worksheet dataframe produced by reading Excel with pandas.r)clsrs rYrHz&_ConnectedComponents.from_worksheet_dfss<   r[list[_ConnectedComponent]cjj\}}tj||fj}jj}d||D}t j||}||t j |}t fd|DS)z=The `_ConnectedComponent` objects comprising this collection.c,g|]}t|Srtuple).0pairs rY z>_ConnectedComponents._connected_components..sKKK45;;KKKr[c:g|]}tj|Sr)rr)rcomponent_node_setrds rYrz>_ConnectedComponents._connected_components..*s7*((:z@_ConnectedComponents._merge_overlapping_tables..?sqwr[)keyN)sortedrrr)rdrsorted_componentscurrent_componentrUs rYrz._ConnectedComponents._merge_overlapping_tables1s##7=N=NOOO * . .I ($-!"3"999$5$;$;I$F$F!!('''$-!!  (# # # # # # ) (r[N)rr)r-r)rrr-r )r-r)rrr-r) r|r}r~rrer classmethodrHrrrrr[rYrGrGs ****0000!!![!   \ 6$$$$$$$$$$$$r[rGceZdZdZddZeddZdd Zdd Zedd Z edd Z eddZ dS)rIaDistinguishes core-table from leading and trailing title rows in a subtable. A *subtable* is a contiguous block of populated cells in the spreadsheet. Leading or trailing rows of that block containing only one populated cell are called "single-cell rows" and are not considered part of the core table. These are each emitted separately as a `Text`-subtype element. rJrc||_dSr^) _subtable)rdrJs rYrez_SubtableParser.__init__as !r[r-pd.DataFrame | Nonect|j}|t|jkrdSt|jt|jz }|j||S)zCThe part between the leading and trailing single-cell rows, if any.N)len _leading_single_cell_row_indicesr!_trailing_single_cell_row_indices)rdcore_table_startcore_table_stops rYrNz_SubtableParser.core_tabledsgtDEE s4>22 2 24 dn--D4Z0[0[[~.>??r[ Iterator[str]c#K|jD]3}|jj|jdV4dS)z8Generate the cell-text for each leading single-cell row.rN)rrrdropnardrow_idxs rYrKz3_SubtableParser.iter_leading_single_cell_rows_textsvsS< @ @G.%g.5577dfd }t|S)zEIndex of each leading single-cell row in subtable, in top-down order.r- Iterator[int]c3ZKtjD]\}}||krdS|VdSr^)r>_single_cell_row_indices next_row_idxidxrds rY$iter_leading_single_cell_row_indicesz^_SubtableParser._leading_single_cell_row_indices..iter_leading_single_cell_row_indicessQ%.t/L%M%M # #! c,&&FF""""" # #r[r-rr)rdrs` rYrz0_SubtableParser._leading_single_cell_row_indicess; # # # # # # 99;;<<dfd }t|S)z=Index of each single-cell row in subtable, in top-down order.r-rc3KtjD]%\}\}}|dkr!|V&dS)Nr!)r>riterrowscount)r_rowrds rYiter_single_cell_row_idxszK_SubtableParser._single_cell_row_indices..iter_single_cell_row_idxss_!*4>+B+B+D+D!E!E   Xa99;;!##   r[rr)rdrs` rYrz(_SubtableParser._single_cell_row_indicess;       ..00111r[ctjtjkrdSdfd }tt t |S)zFIndex of each trailing single-cell row in subtable, in top-down order.rr-rc3Ktjdz }jdddD]}||krdS|V|dz}dS)z... moving from end upward ...r!N)rrrrs rY%iter_trailing_single_cell_row_indicesz`_SubtableParser._trailing_single_cell_row_indices..iter_trailing_single_cell_row_indicesslt~..2L4TTrT: " ",&&FF""""!  " "r[r)rrrrreversedr)rdrs` rYrz1_SubtableParser._trailing_single_cell_row_indicessr t4 5 5T^9L9L L L2 " " " " " "Xd#H#H#J#JKKLLMMMr[N)rJr)r-r)r-r)r-r) r|r}r~rrerrNrKrOrrrrr[rYrIrIXs""""@@@\@"@@@@ @@@@  = = =\ = 2 2 2\ 2NNN\NNNr[rIr:rc4t|rtt|St|rt|St |rt |St |rt|St|S)z=Create `Text`-subtype document element appropriate to `text`.r<) rrr rrrrrrr<s rYrLrLs ]4001111 "4 ( (T"""" #D ) )$'''' 4 $r[rRr8rPrc<t|||j|jS)N)r7r8r'r9)rrDr9)rRr8rPs rYrMrMs- ((    r[r^)r'r(r"r)r#r*r$r*r%r*r&r+r,rr-r.)r:rr-r)rRrr8r+rPr=r-r)9r __future__rrwtempfilertypingrrrrnetworkxrnumpyrpandasrstyping_extensionsr r unstructured.chunkingr unstructured.cleaners.corer unstructured.common.html_tabler unstructured.documents.elementsrrrrrrrunstructured.file_utils.modelr&unstructured.partition.common.metadatarr unstructured.partition.text_typerrrrunstructured.utilsrr__annotations__r XLSXrZr=rrGrIrLrMrr[rYrs#666"""""" ))))))............--------777777444444444444322222YYYYYYYY ,+++++..... "Z!% "& !ZZZZZZz?T?T?T?T?T?T?T?TD6>6>6>6>6>6>6>6>rR$R$R$R$R$R$R$R$jPNPNPNPNPNPNPNPNf    r[