TAin>SrSSKrSSKrSSKrSSKrSSKJr SSKJ r SSKJ r SSKJ r SSKJ r SS KJ r "S S 5r"S S \5rg)zDDefines a base parser. As well as generic methods for other parsers.N)Table) bbox_from_str)compute_accuracy)compute_whitespace)get_table_index) text_in_bboxc\rSrSrSrSSjrSrSrSrSr S r \ S 5r S r S rS rSrSrSrSrg) BaseParserzDefines a base parser.Nc XlX lX0l0UlSUlX@lXPlX`lXplXl SUl SUl SUl 0Ul U (dSUl gg)Ni,)id table_regions table_areastable_bbox_parsescolumns copy_text split_text strip_text shift_text flag_sizerootnamet_bbox resolution parse_details) self parser_idrrrrrrrdebugs P/var/www/html/land-ocr/venv/lib/python3.13/site-packages/camelot/parsers/base.py__init__BaseParser.__init__sl*&!# "$$$"  !%D cJ[URR5SSS9$)zeReturn a list of table bounding boxes sorted by position . Returns ------- [type] [description] c US$)Nxs r)BaseParser.table_bboxes..>s1Q4r"T)keyreverse)sortedrkeysrs r table_bboxesBaseParser.table_bboxes6s#d,,113QUVVr"c xXlXlX lX0lX@lXPlX`lXplURuUlUl [RRUR5uUl n URb3URURS'UR URS'gg)zPrepare the page for parsing.Nrr)filename layout_kwargslayout dimensionspageimageshorizontal_text vertical_text pdf_width pdf_heightospathsplitextrrrr) rr3r5r6page_idxr8r9r:r4__s rprepare_page_parseBaseParser.prepare_page_parse@s! * $  .**.//'GG,,T]]; r    )262D2DD   /040@0@D  } - *r"c/nURcURU5 U$URH)n[[U5U5nURU5 M+ U$)zIf regions have been specified, filter textlines to these regions. Parameters ---------- textlines : list list of textlines to be filtered Returns ------- filtered_textlines : list of textlines within the regions specified )rextendr r)r textlinesfiltered_textlines region_str region_texts r_apply_regions_filter BaseParser._apply_regions_filter[se     %  % %i 0 "!#00 *=+DiP "))+61"!r"cUR(dl[RRUR5nUR (a[ R"US3SS9 g[ R"SU3SS9 gg)zDetect image only documents and warns. Returns ------- has_no_text : bool Whether the document doesn't have any text at all. z8 is image-based, camelot only works on text-based pages.r%) stacklevelzNo tables found on rTF)r9r=r>basenamerr8warningswarn)rrs r_document_has_no_text BaseParser._document_has_no_textqsm##ww'' 6H{{ j!>>  3H:>1Mr"c^[X45nURUlUS-UlX%lU$)aInitialize new table object, ready to be populated. Parameters ---------- table_idx : int Index of this table within the pdf page analyzed bbox : set bounding box of this table within the pdf page analyzed cols : list list of coordinate boundaries tuples (left, right) rows : list list of coordinate boundaries tuples (bottom, top) Returns ------- table : camelot.core.Table r%)rr7order_bbox)r table_idxbboxcolsrowstables r_initialize_new_table BaseParser._initialize_new_tables.&d!YY !m   r"cU$)z Reduces index of a text object if it lies within a spanning cell. Only useful for some parsers (e.g. Lattice), base method is a noop. r&)tidxrs r _reduce_indexBaseParser._reduce_indexs  r"c /nSHnURUHn[UUUURURURS9upV[ U5S:dMBUSSSS:wdMPUR U5 [U5RXURS9nUHupxn XRUUl M M M U$)zCompute parse errors for the table . Parameters ---------- table : camelot.core.Table Returns ------- Tuple Parse errors )vertical horizontal)rrrrNr)re)r) rrrrrlenappendtyper`rcellstext) rrZ pos_errors directionr^indiceserrorr_idxc_idxrjs rcompute_parse_errorsBaseParser.compute_parse_errorss 4I[[+!0#"nn# "w /89y!U1Xy 9  #'#B#B4#S D(($#(WE NN5 ! >:s C c LURUlURUlURUR;aURURUlO[ SURS35 ////4$URUlURU5n[SU//5Ul URbURUR5 URn[R"U5UlUR R"Ul[%U5UlUR(UR*4Ul/nUR/UR0Vs/sH0oUR2UR4UR6UR84PM2 sn5 UR/UR:Vs/sH0oUR2UR4UR6UR84PM2 sn5 XAlUR0UR:-Ulgs snfs snf)*Record data about the origin of the table.zWarning: Bounding box z not found in table_bbox_parses.dN) rflavorr3rUrparseprintrrqraccuracyrcopy_spanning_textdatapd DataFramedfshaper whitespacer;r<pdf_sizerEr9x0y0x1y1r:_textrF)rrZrkrrr^s rrecord_parse_metadata BaseParser.record_parse_metadatasww  ;;$00 000=EK ( 5UV r2r> !"00..u5 )C+<*=> >> %  $ $T^^ 4zz<<%hhnn -d3..$//: 8L8LM8L1ttQTT144.8LMN 8J8JK8J1ttQTT144.8JKL ..1C1CCNKs "7H97H!)rrr6r3rr9rr8r5r4r7rr<r;rrrrrrrrrr:)NNNFrNFF)__name__ __module__ __qualname____firstlineno____doc__r r0rBrJrQr[ staticmethodr`rqrxr}rrr__static_attributes__r&r"rr r st  &DWA6",*2!F$$$8 Dr"r c^\rSrSrSrS U4Sjjr\SSj5r\SSj5r\S5r \S5r \S5r S r S r U4S jrS rU=r$)TextBaseParseriz Base class for all text parsers.c |>[T U]UUUUUUU S9 X@lUR5 XlXlXlg)z:Initialize the text base parser class with default values.)rrrrrrN)superr r_validate_columnsedge_tolrow_tol column_tol)rrrrrrrrrrrrr| __class__s rr TextBaseParser.__init__sN  '#!!       $r"cSn/n/nURSS9 UVs/sH)oUR5R5(dM'UPM+ nnUHpnUc URnOL[R "X%RUS9(d(UR [USS95 /nURnUR U5 Mr UR [USS95 U$s snf)a  Group PDFMiner text objects into rows vertically within a tolerance. Parameters ---------- text : list List of PDFMiner text objects. row_tol : int, optional (default: 2) Returns ------- rows : list Two-dimensional list of text objects grouped into rows. Nc4UR*UR4$rt)rrr's rr),TextBaseParser._group_rows..Ms!$$ r")r+abs_tolcUR$rtrr^s rr)rWsqttr"cUR$rtrrs rr)r]sqttr")sortget_textstriprmathisclosergr-)rjrrow_yrYtempr^non_empty_texts r _group_rowsTextBaseParser._group_rows9s" - .%)BTZZ\-?-?-A!TBA }\\%w? F4^<= KKN  F4^45 !Cs &CCc`/nUGH$nU(dURU5 MUSnUS:alUSUS::d![R"USUSUS9(a,[USUS5n[ USUS5nXe4US'MURU5 MUS:dMUSUS::aj[R"USUS[ U5S9(aURU5 M[USUS5n[ USUS5nXe4US'GMURU5 GM' U$)aMerge column boundaries if they overlap or lie within a tolerance. Parameters ---------- cl : list List of column x-coordinate tuples. column_tol : int, optional (default: 0) Returns ------- merged : list List of merged column x-coordinate tuples. rerr%r)rgrrmaxminabs)clrmergedhigherlower upper_bound lower_bounds r_merge_columnsTextBaseParser._merge_columns`s$ F f%r ?ayE!H, q 58Z1'*%(F1I&> &)%(F1I&> &1%?r  f-!^ayE!H,<<q 58S_U"MM&1*-eAhq *BK*-eAhq *BK*5)CF2J f-/0 r"cUVs/sH'n[SU55[SU55/PM) nn[S[U5S- 5H$nXEnXES-nUSUS-S- =US'US'M& XSS'X$SS'U$s snf)aA Make row coordinates continuous. For the row to "touch" we split the existing gap between them in half. Parameters ---------- rows_grouped : list Two-dimensional list of text objects grouped into rows. text_y_max : int text_y_min : int Returns ------- rows : list List of continuous row y-coordinate tuples. c38# UHoRv M g7frt)r.0r^s r ,TextBaseParser._join_rows..s!q!qc38# UHoRv M g7frt)rrs rrrs&7QttQrrr%rre)rrrangerf) rows_grouped text_y_max text_y_minrrow_boundariesitop_row bottom_rows r _join_rowsTextBaseParser._join_rowss,CO BNQS!q! !3&7Q&7#7 8,  q#n-12A$'G'A.J*1!*z!}*D)I IGAJA3 *q! *r1 s.BczU(a[RXS9nUVs/sHn[U5PM nnUVVs/sH>n[U5[U5:XdMUHoURUR 4PM M@ nnnUR [R[U555 U$s snfs snnf)aGAdds columns to existing list. By taking into account the text that lies outside the current column x-coordinates. Parameters ---------- cols : list List of column x-coordinate tuples. text : list List of PDFMiner text objects. ytol : int Returns ------- cols : list Updated list of column x-coordinate tuples. )r) rrrfrrrrErr-)rXrjrrelementsr^new_colss r _add_columnsTextBaseParser._add_columnss* !--d-DD(,-1AH-&*&*c!fH .E q!qtt q d  KK55fX6FG H .sB2B7%B7cH[U5n[S[U55Vs/sHo0USXS- S-S- PM nnURSU5 UR U5 [S[U5S- 5Vs/sHo0UXS-4PM nnU$s snfs snf)zMakes column coordinates continuous. Parameters ---------- cols : list List of column x-coordinate tuples. text_x_min : int text_y_max : int Returns ------- cols : list Updated list of column x-coordinate tuples. r%rr)r-rrfinsertrg)rX text_x_min text_x_maxrs r _join_columnsTextBaseParser._join_columnss"d|;@CI;NO;Naadq5k!n,1;NO Az" J05aTQ0GH0G1a$1u+&0GH PIs BBcURbFURb8[UR5[UR5:wa [S5eggg)Nz1Length of table_areas and columns should be equal)rrrf ValueErrorr/s rr TextBaseParser._validate_columnssM    'DLL,D4##$DLL(99 !WXX:-E 'r"c lURXX45nUR5nURU5 U$rt)r[ set_all_edgesr)rrVrWrXrYr|rZs rr}TextBaseParser._generate_tables5**9DG##% ""5) r"c2>[TU]U5 SUlg)rN)rr _segments)rrZrs rr$TextBaseParser.record_parse_metadatas %e,r")rrrr) NNNFFr2rrF)r)r)rrrrrr rrrrrrrr}rr __classcell__)rs@rrrs* %>$$L((T>:.Y r"r)rrr=rOpandasrcorerutilsrrrrr r rr&r"rrsCJ !$&# CDCDLZZZr"