Wj]ddlmZddlZddlmZmZmZ ddlm Z n #e $rdZ YnwxYwd,d Z d-d.dZ d/d0dZ d1dZ d2d3dZd4dZd4d Z d5d6d+ZdS)7) annotationsN)ListOptionalTuple)cKDTree node_bboxes np.ndarraymax_gap_distancefloatreturnTuple[np.ndarray, np.ndarray]c  |jd}g}g}|dddf|dddf|dddf|dddff\}}}}||zdz } ||zdz } t|D]C} ||| k} tj| r/tj| d} tjdtj|| || tj|| || z }| |dk}t|dkr|||| z }tj| || | z }|dz|z}|tj |}|tj ||kr,| | |g| d||| k}tj|r/tj|d}tjdtj|||| tj|||| z }||dk}t|dkr|| ||z }tj| || | z }|dz|z}|tj |}|tj ||kr,| | |g| d||| k}tj|r/tj|d}tjdtj|||| tj|||| z }||dk}t|dkr|| ||z }tj| || | z }|dz|z}|tj |}|tj ||kr,| | |g| d||| k}tj|r/tj|d}tjdtj|||| tj|||| z }||dk}t|dkr|||| z }tj| || | z }|dz|z}|tj |}|tj ||kr,| | |g| dEt|dkrBtj gggtj tj gtj fStj |tj j } tj |tj }!| |!fS)Nr@dtype)shaperangenpanywheremaximumminimumlenabsargminappendarrayint64T)"rr Nedges directionsx0y0x1y1cxcyi above_mask above_indiceshorizontal_overlaps valid_abovey_gaps x_distances distances closest_idx below_mask below_indices valid_below left_mask left_indicesvertical_overlaps valid_leftx_gaps y_distances right_mask right_indices valid_right edge_indexedge_directionss" /Users/tedpeters/Documents/Projects/Claude Projects/TPC Houses/house-viewer/.venv/lib/python3.11/site-packages/pymupdf/layout/presidio/graph.pyfind_directional_neighborsrD s!A EJ A& AAAqD(9;qqq!t;LkZ[Z[Z[]^Z^N__NBB r'SB r'SB 1XX?)?)2a5j 6*   )HZ003M"$*2:b/A77"*R EVXZ[\X]:^:^^## ((;a(?@K;!##K2a50 fR _r!u%<== "SL;6 )")I*>*>? RYy1126FFFLL![!1222%%a(((2a5j 6*   )HZ003M"$*2:b/A77"*R EVXZ[\X]:^:^^## ((;a(?@K;!##AK0 fR _r!u%<== "SL;6 )")I*>*>? RYy1126FFFLL![!1222%%a((("Q%Z 6)   )8I..q1L " 2:b.166B|DTVXYZV[9\9\\!! &&7!&; RYy1126FFFLL![!1222%%a(((2a5j 6*   )HZ003M " 2:b/A77"*R EVXZ[\X]:^:^^!! ((9A(=>K;!##K2a50 fR _r!u%<== "SL;6 )")I*>*>? RYy1126FFFLL![!1222%%a((( 5zzQxR11128Bbh3O3O3OOO%rx0002Jhz:::O  &&@ tolerance max_distanceOptional[float] page_width page_heightch|jd}|dkr"tjgggtjS|dddf|dddf|dddf|dddff\}}}} ||zdz } || zdz } g} t |D]_} t | dz|D]G}|D| || | z }| || | z }tj||z||zz|krIt || ||z |kpCt || ||z |kp!t | | | |z |k}t || ||z |kpCt | | | |z |kp!t | | | |z |k}|s|r.| | |g| || gIa| s"tjgggtjStj| tjjS)Nrrrrrr) rrr r!rsqrtrrr")rrGrHrJrKr#r&r'r(r)r*r+r$r,jdxdy aligned_x aligned_ys rCfind_alignment_edgesrS^sl !AAvvxR1111 A& AAAqD(9;qqq!t;LkZ[Z[Z[]^Z^N__NBB r'SB r'SB E 1XX%%q1ua % %A'URU]URU]727R"W,-- <<11 &&)372a52a5=))Y672a52a5=))Y6  11 &&)372a52a5=))Y672a52a5=))Y6   %I % aV$$$ aV$$$) %, 2xR1111 8E * * * ,,rEkintc|jd}|dks|dkr"tjgggtjS|dddf|dddfzdz }|dddf|dddfzdz }tj||gd}t t |}t |dz|}|||\}} | jdkr"tjgggtjS| ddddf} | jd} tj tj || } | } tj| | gd tjSg}tj |} t|D]}|||z }|dddfdz|dddfdzz}tj||<tj| |f}|dt ||dz }|D]&}||t#|g'|s"tjgggtjStj|tjjS) Nrrrrrraxis)rU)rrr r!stackrminqueryndimrepeatarangeflattenastyperinflexsortrrVr")rrUr#r*r+centerstreek_query_indicesneighbor_indicesk_actualsrcdstr$r,diffsd2ordernn_idxrNs rCfind_knn_edgesrqs!AAvvaxR1111 aaad k!!!Q$/ /3 6B aaad k!!!Q$/ /3 6BhBxa(((Gwa!eQ--ZZ7Z33 7 <1  8RHBH555 5"111abb5>#)!,i ! h//&&((xc +++2228<<< EillG 1XX&&'!*$ 111a4[A aaad q 0 01 GR=))QA' & &A LL!SVV % % % % & 2xR1111 8E * * * ,,rErAc |jd}|dkr tjdtjSd||d}||d}|dddf|dddfz }|dddf|dddfz }|dddf|dddfz }|dddf|dddfz }tj|dddf|dddf} tj|dddf|dddf} tj|dddf|dddf} tj|dddf|dddf} | | z } | | z }fd }||}||}||}||}|| }||}g}||dddf|dddfz |z ||dddf|dddfz |z ||dddf|dddfz |z ||dddf|dddfz |z |tj||z |tj||z ||dddf| z |z ||dddf| z |z || |dddfz |z || |dddfz |z |tj||z |tj||z ||dddf| z |z ||dddf| z |z || |dddfz |z || |dddfz |z |tj||z |tj||z tj|d  tj}tj |d d d S)am Compute SCE-style 18-d edge features (relative positions, scale, enclosing box). See external/sce train/infer/common_util get_edge_transform_bbox and docs/EDGE_FEATURES.md. S = source (edge_index[0]), O = target (edge_index[1]). R = enclosing bbox of S and O. Bboxes [x0, y0, x1, y1] (same as Presidio). Returns: [E, 18] float32. rr)rrư>Nrrc8tj|dk|S)Nr)rr)xepss rCz.compute_sce_edge_features_18..sRXa1fc1--rErXggj@gj)nanposinfneginf) rrzerosfloat32rrrlogrZra nan_to_num)rrAESOswshowohR_x1R_y1R_x2R_y2rwrhsafesw_sh_ow_oh_rw_rh_foutrws @rCcompute_sce_edge_features_18rs AAvvxrz2222 CJqM"AJqM"A 111a41QQQT7 B 111a41QQQT7 B 111a41QQQT7 B 111a41QQQT7 B :a1gqAw ' 'D :a1gqAw ' 'D :a1gqAw ' 'D :a1gqAw ' 'D B B - - - -D $r((C $r((C $r((C $r((C $r((C $r((C AHHa1g!!!Q$3 &'''HHa1g!!!Q$3 &'''HHa1g!!!Q$3 &'''HHa1g!!!Q$3 &'''HHRVC#I  HHRVC#I  HHa1gn #$$$HHa1gn #$$$HHdQqqq!tWn #$$$HHdQqqq!tWn #$$$HHRVC#I  HHRVC#I  HHa1gn #$$$HHa1gn #$$$HHdQqqq!tWn #$$$HHdQqqq!tWn #$$$HHRVC#I  HHRVC#I   (11    $ $RZ 0 0C =#c$ ? ? ??rErBOptional[np.ndarray]node_font_sizesnode_draw_ordersc|jd}|dkr tjdtjS|dddf|dddf|dddf|dddff\}} } } || zdz } | | zdz } | |z }| | z }|d}|d}||| || || |f\}}}}||| || || |f\}}}}| || |}}| || |}}||||}}||||}!} ||z |z }"||z |z }#tj|tj}$||k}%||k}&||%||%z |$|%<||&||&z |$|&<tj|tj}'||k}(||k})||(||(z |'|(<||)||)z |'|)<tjdtj||tj||z }*tjdtj||tj||z }+|*tjtj|| d z },|+tjtj||!d z }-|*|+z}.||z}/| |!z}0|/|0z|.z }1|.tj|1d z }2|.tj|/d z }3|.tj|0d z }4tjtj| tj|d z d }5tjtj|!tj|d z d }6tj|d ftj}7|6t||kr#|dk|dkz}8d |7|8||8f<d |7|8d f<ntj|"}9tj|#}:|9|:k};|9|:k}<|;|<z}=|;|#dkz}>|;|#dkz}?|<|"dkz}@|<|"dkz}Ad |7|?df<d |7|>df<d |7|Adf<d |7|@df<d |7|=d f<||z |z }B||z |z }C||z |z }D||z |z }E||z |z }F||z |z }Gtj |"|#|$|z |'|z |,|-|5|6|B|C|D|E|F|G|2|3|4gd  tj}Htj |H|7gd }I|(t|dkr||}J||}Kd }Ld }Mtj|J|M}Ntj|K|M}O|Ntj|O|Lz }Pd}Qd}Rtj |P|R|Q}Ptj|P}Stj|S}T|J|Mk|Jdkz}U|K|Mk|Kdkz}V|U|Vz tj}W|S|Wz}S|T|Wz}Ttj |S|Tgd  tj}Xn"tj|dftj}Xtj |I|Xgd }I|Bt|dkr/||}Y||}Z|Z|Yz }[tj |[dddz }[n tj|tj}[|[ dd}\tj |I|\gd }It||}]tj |I|]gd }I|IS)u Compute edge features for each edge (vectorized). Appends SCE 18-d features (see compute_sce_edge_features_18) ¡æ [E, 43]. Edge features (25-d base): - normalized_dx = (cx_j - cx_i) / page_width - normalized_dy = (cy_j - cy_i) / page_height - signed_gap_x (horizontal gap) - signed_gap_y (vertical gap) - x_overlap_ratio - y_overlap_ratio - log(w_j / w_i) (relative width ratio) - log(h_j / h_i) (relative height ratio) - delta_left = (x0_j - x0_i) / page_width (left-to-left delta) - delta_right = (x1_j - x1_i) / page_width (right-to-right delta) - delta_center_horizontal = (cx_j - cx_i) / page_width (horizontal center delta) - delta_top = (y0_j - y0_i) / page_height (top edge delta) - delta_bottom = (y1_j - y1_i) / page_height (bottom edge delta) - delta_middle = (cy_j - cy_i) / page_height (vertical center/middle delta) - iou = intersection_area / union_area (Intersection over Union - 2D overlap) - ioa_i = intersection_area / area_i (Intersection over Area of node i) - ioa_j = intersection_area / area_j (Intersection over Area of node j) - direction one-hot: {up, down, left, right, other} (5 dims) - font_log_ratio = log(font_size_u / font_size_v) (font size contrast) - font_abs_log_ratio = |log(font_size_u / font_size_v)| (absolute font size contrast) - delta_draw_order = draw_order_j - draw_order_i (difference of normalized draw orders) Args: node_bboxes: Array of shape [N, 4] - [x0, y0, x1, y1] edge_index: Array of shape [2, E] - edge indices page_width: Page width in points page_height: Page height in points edge_directions: Optional array of shape [E] - direction labels (0=up, 1=down, 2=left, 3=right) node_font_sizes: Optional array of shape [N] - font sizes per node (font_size_median or font_size) node_draw_orders: Optional array of shape [N] - draw order indices (0-indexed) per node Returns: Edge feature matrix [E, 43]. rrr+rNrrrrtg?rXg$@g?g@@i )rrr|r}rrr~rrrZra concatenateclipreshaper)^rrArJrKrBrrrr&r'r(r)r*r+whrkrlx0_iy0_ix1_iy1_ix0_jy0_jx1_jy1_jcx_icy_icx_jcy_jw_ih_iw_jh_jrOrPgap_xleft_ofright_ofgap_yabovebelow x_overlap y_overlapx_overlap_ratioy_overlap_ratiointersection_areaarea_iarea_j union_areaiouioa_iioa_j log_w_ratio log_h_ratiodirection_onehot valid_dirsabs_dxabs_dyvertical horizontaldiagonaldownuprightleft delta_left delta_rightdelta_center_horizontal delta_top delta_bottom delta_middlefeatures edge_features font_size_i font_size_jEPS MIN_FONT_SIZEfont_size_i_safefont_size_j_safe font_ratio MAX_RATIO MIN_RATIOfont_log_ratiofont_abs_log_ratiovalid_ivalid_j font_valid font_features draw_order_i draw_order_jdelta_draw_orderdraw_order_featuresces^ rCcompute_edge_featuresrs` AAvvxrz2222!A& AAAqD(9;qqq!t;LkZ[Z[Z[]^Z^N__NBB r'SB r'SB RA RA Q-C Q-C Wbgr#w3?D$dWbgr#w3?D$dC"S'$DC"S'$DvqvCvqvC + #B + $B HQbj ) ) )ETkGd{H']T']2E'NXh78E(O HQbj ) ) )E 4KE 4KE;e,E%L%[4;./E%L 1bjt44rz$7M7MMNNI 1bjt44rz$7M7MMNNI"*RZS-A-A4"H"HHO"*RZS-A-A4"H"HHO"I- 3YF 3YF&#44J bjT:: :C  64 8 8 8E  64 8 8 8E&C"*S$*?*?$?FFGGK&C"*S$*?*?$?FFGGKxAbj999"s?';';q'@'@%*!/CD DG_Z%@@A+.*a((F?f_ 9 {*26" q !b1f%R1W%"%Q$'q!$'q!%("(+1%++J$;*,K $d{j8 +I4K;.L4K;.Lx B     !" #"vbj!!# (NH.>#?aHHHM"s?';';a'?'?%c* %c*  :k=AA:k=AA& 3CS(I(II   WZI>>  ++ VN33-/K&4HI-/K&4HI'// ;; (*4/*<  " 6"*%%  !Qrz::: NM=#AJJJM#,<(=(=(A(A', ', ',67#3S"==D8ARZ888)11"a88NM3E#FQOOOM &{J ? ?CNM3#7a@@@M rE List[str]c gdS)z[Names for SCE 18-d edge features (see compute_sce_edge_features_18, docs/EDGE_FEATURES.md).)sce_src_tgt_dx_w1sce_src_tgt_dy_h1sce_tgt_src_dx_w2sce_tgt_src_dy_h2 sce_log_w1_w2 sce_log_h1_h2sce_src_left_r sce_src_top_rsce_r_left_src sce_r_top_src sce_log_w1_rw sce_log_h1_rhsce_tgt_left_r sce_tgt_top_rsce_r_left_tgt sce_r_top_tgt sce_log_w2_rw sce_log_h2_rhrrErCget_sce_edge_feature_namesrs   rEc(gdtzS)zUNames of edge features in compute_edge_features output order (43-d, includes SCE 18).)rOrP gap_x_norm gap_y_normrrrrrrrrrrrrr direction_updirection_downdirection_leftdirection_rightdirection_diagonalrrr)rrrErCget_edge_feature_namesrs&    #$$ %%rE@@FList[List[float]] k_nearestalignment_toleranceuse_alignment_edgesboolOptional[List[float]]Optional[List[int]])Tuple[np.ndarray, np.ndarray, np.ndarray]c b tj|tj} | jd} | dkratjgggtjtjdtjtjgtjfS| dkrdtjdgdggtjtjdtjtjdgtjfSig} t | |\} } | jddkrt| jdD]g}t| d|ft| d|f}}||f}|vr/t| ||<| ||ght| |}|jddkrot|jdD]T}t|d|ft|d|f}}||f}|vrd|<| ||gU|rtj |dz|dzzdz}t| |||| }|jddkrot|jdD]T}t|d|ft|d|f}}||f}|vrd|<| ||gUt| dkr| d krOt| D]>}t| D],}||kr$||f}|vrd|<| ||g-?nt| t|| dz }|jddkrot|jdD]T}t|d|ft|d|f}}||f}|vrd|<| ||gU| }fd |D}tj|tjj}tj|tj}| tj|tjnd}| tj|tjnd}t!| |||||| }|||fS) Nrrrr)rrrrg?)rHrJrKrc*g|]\}}||fSrr).0r,rNall_edges_dicts rC zbuild_graph..Rs&CCCAnaV,CCCrE)rBrr)rr r}rr!r|rDrrVrrqrMrSrr[r"r)rrJrKrrr rrrnode_bboxes_arrr#all_edges_listnn_edges nn_directionser,rNedge_key knn_edgesmax_alignment_distance align_edges all_edgesall_directionsrArBnode_font_sizes_arrnode_draw_orders_arrrrs @rC build_graphr+s h{"*===Oa AAvv Hb"XRX . . . HWBJ / / / HRrx ( ( (   Avv HqcA3Zrx 0 0 0 HWBJ / / / HaS ) ) )  NN8JZ[[Hm~a1x~a()) . .Ax1~&&HQTN(;(;qA1vH~--+.}Q/?+@+@x(%%q!f--- ::IqAyq)** . .AyA''Yq!t_)=)=qA1vH~--+,x(%%q!f---2!#q;!;K)K!L!Ls!R*  /!#      Q ! # #;,Q/00 2 2;q!t,--s;q!t3D/E/E1q6>11/0N8,"))1a&111 >a 771XX : :q::AAvv$%q6#>9978N84*111a&999 : :'Iq1u8M8MNNIq!A%%yq12266AyA//Yq!t_1E1EqA !1vH~5534x0&--q!f555ICCCCCCCN)284446Jh~RX>>>OIXId"(?"*EEEEjnK[Kg28$4BJGGGGmq)'+-M }o 55rE)rr r r r r )rFNNN) rr rGr rHrIrJrIrKrIr r )rT)rr rUrVr r )rr rAr r r )NNN)rr rAr rJr rKr rBrrrrrr r )r r)rTrFrFNN)rrrJr rKr rrVrr r r rrrrrrr r) __future__rnumpyrtypingrrr scipy.spatialr ExceptionrDrSrqrrrrr+rrErCr1s""""""((((((((((%%%%%%%GGGO'O'O'O'h(,&*'+ )-)-)-)-)-X!-!-!-!-!-H?@?@?@?@N150415pppppf     % % % %"%("'$)1504h6h6h6h6h6h6h6s ''