o f_@sddlZddlZddlmZmZmZmZmZmZm Z m Z m Z m Z m Z mZddlmZddlmZddlmZddlmZddlmZdd lmZmZdd lmZdd lmZdd lmZdd lmZddlmZddlm Z ddlm!Z!ddlm"Z"ddlm#Z#ddlm$Z$ddlm%Z%e&e'Z(GdddZ)GdddZ*GdddZ+GdddZ,Gddde+Z-Gdd d e-Z.Gd!d"d"e.Z/Gd#d$d$e.Z0Gd%d&d&e-Z1Gd'd(d(e+e,Z2Gd)d*d*e-e,Z3e d+e+d,Z4Gd-d.d.e-ee4Z5Gd/d0d0e5e4Z6Gd1d2d2e6e4e,Z7e e3e2fZ8Gd3d4d4e7e8Z9Gd5d6d6e9Z:Gd7d8d8e9Z;Gd9d:d:e7e9Zd>e<Z>e er?r@rArBrCrD _validate)r(r>r?r@rArBrCrDr)r)r*r+Ps  zLAParams.__init__cCs\|jdur*d}t|jtst|jtst|d|jkr%dks,t|t|dSdS)Nz@LAParam boxes_flow should be None, or a number between -1 and +1r)rBr0r9float TypeError ValueError)r(boxes_flow_err_msgr)r)r*rEds   zLAParams._validatecCsd|j|j|j|jfS)NzM)r?r@rArDr(r)r)r*__repr__pszLAParams.__repr__)r;r<r;r=r;FFr$N) r6r7r8__doc__rGrboolr+rEstrrLr)r)r)r*r:4s8  r:c@s"eZdZdZdeddfddZdS)r.z)Interface for things that can be analyzedlaparamsr$NcCdS)zPerform the layout analysis.Nr)r(rQr)r)r*analyze{zLTItem.analyze)r6r7r8rNr:rTr)r)r)r*r.xsr.c@s,eZdZdZdefddZdefddZdS)LTTextz#Interface for things that have textr$cCsd|jj|fS)Nz<%s %r>) __class__r6get_textrKr)r)r*rLzLTText.__repr__cCt)zText contained in this objectNotImplementedErrorrKr)r)r*rXrUzLTText.get_textN)r6r7r8rNrPrLrXr)r)r)r*rVsrVc@seZdZdZdeddfddZdefddZd ede fd d Z d ede fd d Z d ede fddZ d ede fddZ deddfddZde fddZddde fddZdddefddZdddefddZddde fddZdddefdd Zdddefd!d"ZdS)# LTComponentzObject with a bounding boxbboxr$NcCst|||dSr&)r.r+set_bboxr(r^r)r)r*r+s zLTComponent.__init__cCsd|jjt|jfS)Nz<%s %s>)rWr6rr^rKr)r)r*rLzLTComponent.__repr___cCrZr&rIr(rbr)r)r*__lt__zLTComponent.__lt__cCrZr&rcrdr)r)r*__le__rfzLTComponent.__le__cCrZr&rcrdr)r)r*__gt__rfzLTComponent.__gt__cCrZr&rcrdr)r)r*__ge__rfzLTComponent.__ge__cCsB|\}}}}||_||_||_||_|||_|||_||_dSr&)x0y0x1y1widthheightr^)r(r^rjrkrlrmr)r)r*r_s    zLTComponent.set_bboxcCs|jdkp |jdkSNr)rnrorKr)r)r*is_emptyrYzLTComponent.is_emptyr-cC2t|ts Jtt||j|jko|j|jkSr&)r0r]rPtyperjrlr(r-r)r)r* is_hoverlapzLTComponent.is_hoverlapcCJt|ts Jtt|||rdStt|j|jt|j|jSrp r0r]rPrsruminabsrjrlrtr)r)r* hdistance "zLTComponent.hdistancecCJt|ts Jtt|||r#tt|j|jt|j|jSdSrprxrtr)r)r*hoverlap "zLTComponent.hoverlapcCrrr&)r0r]rPrsrkrmrtr)r)r* is_voverlaprvzLTComponent.is_voverlapcCrwrp r0r]rPrsrryrzrkrmrtr)r)r* vdistancer|zLTComponent.vdistancecCr}rprrtr)r)r*voverlaprzLTComponent.voverlap)r6r7r8rNrr+rPrLobjectrOrergrhrir_rqrurGr{r~rrrr)r)r)r*r]s  r]c@seZdZdZ       ddedeedededed ee d ee d eee d ee e e fd dfddZ d efddZdS)LTCurvez A generic Bezier curve The parameter `original_path` contains the original pathing information from the pdf (e.g. for reconstructing Bezier Curves). `dashing_style` contains the Dashing information if any. FN linewidthptsstrokefillevenoddstroking_colornon_stroking_color original_path dashing_styler$c CsJt|t|||_||_||_||_||_||_||_ ||_ | |_ dSr&) r]r+rrrrrrrrrr) r(rrrrrrrrrr)r)r*r+s  zLTCurve.__init__cCsddd|jDS)N,css|]}d|VqdS)z %.3f,%.3fNr)).0pr)r)r* sz"LTCurve.get_pts..)joinrrKr)r)r*get_ptsrazLTCurve.get_ptsFFFNNNN)r6r7r8rNrGrrrOrrrr rr+rPrr)r)r)r*rs>    rc@sreZdZdZ       ddededededed ed eed eed ee e d ee e e fddfddZ dS)LTLinezOA single straight line. Could be used for separating text or figures. FNrp0p1rrrrrrrr$c Cs$t||||g|||||| | dSr&rr+) r(rrrrrrrrrrr)r)r*r+s zLTLine.__init__r)r6r7r8rNrGrrOrrrrr rr+r)r)r)r*rs@     rc@sneZdZdZ       ddededededed eed eed ee e d ee e e fd dfddZ dS)LTRectzMA rectangle. Could be used for framing another pictures or figures. FNrr^rrrrrrrr$c  CsD|\} } } } t||| | f| | f| | f| | fg||||||| dSr&r)r(rr^rrrrrrrrjrkrlrmr)r)r*r+ s zLTRect.__init__r)r6r7r8rNrGrrOrrrrr rr+r)r)r)r*rs<    rc@s8eZdZdZdedededdfddZdefd d ZdS) LTImagezKAn image object. Embedded images can be in JPEG, Bitmap or JBIG2. namestreamr^r$NcCsrt||||_||_|d|df|_|d|_|dd|_|d|_t |jt s7|jg|_dSdS)N)WWidth)HHeight)IM ImageMask)BPCBitsPerComponentr)CS ColorSpace) r]r+rrget_anysrcsize imagemaskbits colorspacer0list)r(rrr^r)r)r*r+As    zLTImage.__init__cCd|jj|jt|j|jfSNz<%s(%s) %s %r>)rWr6rrr^rrKr)r)r*rLL zLTImage.__repr__) r6r7r8rNrPrrr+rLr)r)r)r*r;s rc@s0eZdZdZdeddfddZdefddZdS) LTAnnoaActual letter in the text as a Unicode string. Note that, while a LTChar object has actual boundaries, LTAnno objects does not, as these are "virtual" characters, inserted by a layout analyzer according to the relationship between two characters (e.g. a space). textr$NcCr%r&_text)r(rr)r)r*r+]szLTAnno.__init__cC|jSr&rrKr)r)r*rXazLTAnno.get_text)r6r7r8rNrPr+rXr)r)r)r*rUsrc@seZdZdZdededededededed eee e eeffd e d e d d fddZ d efddZd efddZded efddZd S)LTCharz.Actual letter in the text as a Unicode string.matrixfontfontsizescalingriser textwidthtextdispncs graphicstater$Nc Cs~t|||_||_|j|_| |_| |_||||_|rVt |t s'J|\} } | dur4|d} n| |d} d| |d} | | ||jf} | || |f}n| |}d||f} |j|||f}|j\}}}}}}d|||ko||dk|_ t |j| \}}t |j|\}}||kr||}}||kr||}}t|||||f|r|j|_dS|j|_dS)Nr;gMbP?ir)rVr+rrfontnamerradv is_verticalr0tuple get_descentuprightrr]rnsizero)r(rrrrrrrrrrvxvybbox_lower_leftbbox_upper_rightdescentabcdefrjrkrlrmr)r)r*r+hsB       zLTChar.__init__cCs,d|jjt|jt|j|j|j|fS)Nz(<%s %s matrix=%s font=%r adv=%s text=%r>) rWr6rr^r rrrrXrKr)r)r*rLszLTChar.__repr__cCrr&rrKr)r)r*rXrzLTChar.get_textr-cCrR)z.)rrKr)r)r*rXs zLTTextContainer.get_textrM)r6r7r8r+rPrXr)r)r)r*rs rcszeZdZdZdeddffdd ZdefddZd eddfd d Z d e e d ede dfddZ deffdd ZZS) LTTextLinezContains a list of LTChar objects that represent a single text line. The characters are aligned either horizontally or vertically, depending on the text's writing mode. rAr$Ncst||_dSr&)superr+rAr(rArWr)r*r+ zLTTextLine.__init__cCsd|jjt|j|fS)Nz <%s %s %r>)rWr6rr^rXrKr)r)r*rLs zLTTextLine.__repr__rQcCs*|jD]}||qt|tddS)N )rrTrrrrr)r)r*rTs  zLTTextLine.analyzeplaneratiocCrZr&r[)r(rrr)r)r*find_neighbors szLTTextLine.find_neighborscstp |Sr&)rrqrXisspacerKrr)r*rqrazLTTextLine.is_empty)r6r7r8rNrGr+rPrLr:rTrrrrrOrq __classcell__r)r)rr*rs rcseZdZdeddfddZdeddffdd Zd eed ede e fd d Z ddedede fddZ ddedede fddZ ddedede fddZddedede fddZZS)LTTextLineHorizontalrAr$NcCst||t |_dSr&)rr+r_x1rr)r)r*r+ zLTTextLineHorizontal.__init__r-cs\t|tr"|jr"|jt|j|j}|j|j|kr"t |t d|j |_t |dSN ) r0rrArrnrorrjrrrrlrr(r-marginrr)r*r zLTTextLineHorizontal.addrrcs@|j|jjjjf}fdd|DS)aK Finds neighboring LTTextLineHorizontals in the plane. Returns a list of other LTTestLineHorizontals in the plane which are close to self. "Close" can be controlled by ratio. The returned objects will be the same height as self, and also either left-, right-, or centrally-aligned. cRg|]%}t|trj|drj|ds%j|ds%j|dr|qS) tolerance)r0r_is_same_height_as_is_left_aligned_with_is_right_aligned_with_is_centrally_aligned_withrrr(r)r* 1    z7LTTextLineHorizontal.find_neighbors..)rofindrjrkrlrmr(rrrr)rr*r$ " z#LTTextLineHorizontal.find_neighborsrotherrcCt|j|j|kS)zN Whether the left-hand edge of `other` is within `tolerance`. )rzrjr(r rr)r)r*r?z*LTTextLineHorizontal._is_left_aligned_withcCr )zO Whether the right-hand edge of `other` is within `tolerance`. )rzrlr r)r)r*rEr z+LTTextLineHorizontal._is_right_aligned_withcC(t|j|jd|j|jd|kS)zQ Whether the horizontal center of `other` is within `tolerance`. )rzrjrlr r)r)r*rK(z/LTTextLineHorizontal._is_centrally_aligned_withcCr r&)rzror r)r)r*rSrYz'LTTextLineHorizontal._is_same_height_asr5)r6r7r8rGr+r]rrrrrrrOrrrrrr)r)rr*rs*   rcseZdZdeddfddZdeddffdd Zd eed ede e fd d Z ddedede fddZ ddedede fddZ ddedede fddZdedede fddZZS)LTTextLineVerticalrAr$NcCst||t |_dSr&)rr+r_y0rr)r)r*r+XrzLTTextLineVertical.__init__r-cs\t|tr"|jr"|jt|j|j}|j||jkr"t |t d|j |_t |dSr) r0rrArrnrormrrrrrkrrrr)r*r_rzLTTextLineVertical.addrrcs@|j|jjjjf}fdd|DS)aG Finds neighboring LTTextLineVerticals in the plane. Returns a list of other LTTextLineVerticals in the plane which are close to self. "Close" can be controlled by ratio. The returned objects will be the same width as self, and also either upper-, lower-, or centrally-aligned. crr)r0r_is_same_width_as_is_lower_aligned_with_is_upper_aligned_withrrrr)r*rurz5LTTextLineVertical.find_neighbors..)rnrrjrkrlrmrr)rr*rhrz!LTTextLineVertical.find_neighborsrr rcCr )zJ Whether the lower edge of `other` is within `tolerance`. )rzrkr r)r)r*rr z)LTTextLineVertical._is_lower_aligned_withcCr )zJ Whether the upper edge of `other` is within `tolerance`. )rzrmr r)r)r*rr z)LTTextLineVertical._is_upper_aligned_withcCr )zO Whether the vertical center of `other` is within `tolerance`. r)rzrkrmr r)r)r*rrz-LTTextLineVertical._is_centrally_aligned_withcCr r&)rzrnr r)r)r*rrYz$LTTextLineVertical._is_same_width_asr5)r6r7r8rGr+r]rrrrrrrOrrrrrr)r)rr*rWs*   rc@s6eZdZdZd ddZdefddZdefdd ZdS) r1zRepresents a group of text chunks in a rectangular area. Note that this box is created by geometric analysis and does not necessarily represents a logical boundary of the text. It contains a list of LTTextLine objects. r$NcCst|d|_dS)NrF)rr+r#rKr)r)r*r+rzLTTextBox.__init__cCs d|jj|jt|j|fSr)rWr6r#rr^rXrKr)r)r*rLs zLTTextBox.__repr__cCrZr&r[rKr)r)r*get_writing_moderfzLTTextBox.get_writing_moderM)r6r7r8rNr+rPrLrr)r)r)r*r1s  r1c4eZdZdeddffdd ZdefddZZS)LTTextBoxHorizontalrQr$Nc"t||jjddddS)NcS|j Sr&)rmr-r)r)r*z-LTTextBoxHorizontal.analyze..keyrrTrsortrSrr)r*rT zLTTextBoxHorizontal.analyzecCrR)Nzlr-tbr)rKr)r)r*rrfz$LTTextBoxHorizontal.get_writing_moder6r7r8r:rTrPrrr)r)rr*rrcr)LTTextBoxVerticalrQr$Ncr)NcSrr&)rlrr)r)r*rrz+LTTextBoxVertical.analyze..rrrSrr)r*rTr!zLTTextBoxVertical.analyzecCrR)Nztb-rlr)rKr)r)r*rrfz"LTTextBoxVertical.get_writing_moder"r)r)rr*r$r#r$r2cs*eZdZdeeddffdd ZZS)r2rr$Ncst||dSr&)rr+r)r(rrr)r*r+rzLTTextGroup.__init__)r6r7r8rTextGroupElementr+rr)r)rr*r2s"c&eZdZdeddffdd ZZS)LTTextGroupLRTBrQr$Nc:t||jdus J|j|jjfddddS)Ncs"d|jd|j|jSr/)rjrkrmrrBr)r*rs z)LTTextGroupLRTB.analyze..rrrTrBrr rSrr)r*rT  zLTTextGroupLRTB.analyzer6r7r8r:rTrr)r)rr*r'r'cr&)LTTextGroupTBRLrQr$Ncr()Ncs$d |j|jd|jSr/)rjrlrmrr)r)r*rs z)LTTextGroupTBRL.analyze..rr*rSrr)r*rTr+zLTTextGroupTBRL.analyzer,r)r)rr*r.r-r.c@seZdZdeddfddZdedeedee fdd Z ded ee dee fd d Z ded e e deefddZdeddfddZdS)LTLayoutContainerr^r$NcCst||d|_dSr&)rr+groupsr`r)r)r*r+rzLTLayoutContainer.__init__rQrccsd}d}|D]}|dur||o4||o4t|j|j|j||ko4||t|j|j|j k}|j o_||o_| |o_t|j|j|j| |ko_| |t|j|j|j k}|rgt|tsn|rtt|trt||nA|dur~|Vd}n7|r|st|j}||||n#|r|st|j}||||nt|j}|||Vd}|}q|durt|j}|dusJ|||VdSr&)rrryror>rr{rrnr?rCrur~rr0rrrrA)r(rQrobj0lineobj1halignvalignr)r)r* group_objectssr              zLTLayoutContainer.group_objectslinesc cst|j}||i}|D]>}|||j}|g}|D]}||||vr0|||qt|tr:t } nt } t |D] } | | | || <qAqt } |D]}||vrZqS||} | | vrcqS| | | so| VqSdS)z$Group neighboring lines to textboxesN)rr^rrr@rpopr0rrr$r!rsetrq) r(rQr7rboxesr2 neighborsmembersr3boxr-doner)r)r*group_textlinesCs>        z!LTLayoutContainer.group_textlinesr:c stttf}t|jdtdtdtfdd}d|d|dt|ffdd }g}tt |D](}||}t|dt |D]} || } | d ||| t |t | || fq.distcs\t|j|j}t|j|j}t|j|j}t|j|j}t||||f}|||fS)z8Check if there's any other object between obj1 and obj2.) ryrjrkrrlrmr9r difference)r3r@rjrkrlrmrrr)r*isanys z0LTLayoutContainer.group_textboxes..isanyrFrTcss|]}tt|VqdSr&)r r2)rgr)r)r*rsz4LTLayoutContainer.group_textboxes..)r r1r2rr^r]rGr rangerridheapqheapifyrr9heappopheappushr0r$r.r'removeupdaterr)r(rQr:ElementTrArDdistsibox1jbox2r> skip_isanyrid1id2r3r@groupr r)rCr*group_textboxesdsH   &        z!LTLayoutContainer.group_textboxesc Cs&tdd|\}}|D]}||q |sdSt|||}tdd|\}}|D]}||q*t|||}|jdur^|D]}||qAdtdttt t ffdd} |j | dn"| |||_ t } |j D] } | || | qk|j d ddttt||ttt||_dS) NcSs t|tSr&)r0rrr)r)r*rs z+LTLayoutContainer.analyze..cSs|Sr&)rqrr)r)r*rrr=r$cSs,t|trd|j |j fSd|j |jfS)Nrr)r0r$rlrkrjr=r)r)r*getkeys z)LTLayoutContainer.analyze..getkeyrcSrr&r'rYr)r)r*rs)rrTrr6r?rBr1r r9rGr rXr0r"r3r rr]r) r(rQtextobjs otherobjsr- textlinesempties textboxestextboxrZassignerrWr)r)r*rTs8         zLTLayoutContainer.analyze)r6r7r8rr+r:rr]rrr6r1r?rrr2rXrTr)r)r)r*r/s0 P ! [r/c@sJeZdZdZdedededdfddZdefd d Zd e ddfd d Z dS)LTFigurezRepresents an area used by PDF Form objects. PDF Forms can be used to present figures or pictures by embedding yet another PDF document within a page. Note that LTFigure objects can appear recursively. rr^rr$Nc sj||_|_|\}}}}||f|||f|||f||||ff}tfdd|D}t||dS)Nc3s"|] \}}t||fVqdSr&r)rrqrr)r*rs z$LTFigure.__init__..)rrrr/r+) r(rr^rr4ywhboundsr)rdr*r+s , zLTFigure.__init__cCs"d|jj|jt|jt|jfS)Nz<%s(%s) %s matrix=%s>)rWr6rrr^r rrKr)r)r*rLs zLTFigure.__repr__rQcCs|jsdSt||dSr&)rDr/rTrSr)r)r*rTs zLTFigure.analyze) r6r7r8rNrPrrr+rLr:rTr)r)r)r*rbs  rbc @s:eZdZdZd dedededdfdd Zdefd d Z dS) LTPagezRepresents an entire page. Like any other LTLayoutContainer, an LTPage can be iterated to obtain child objects like LTTextBox, LTFigure, LTImage, LTRect, LTCurve and LTLine. rpageidr^rotater$NcCst||||_||_dSr&)r/r+rjrk)r(rjr^rkr)r)r*r+ s zLTPage.__init__cCr)Nz<%s(%r) %s rotate=%r>)rWr6rjrr^rkrKr)r)r*rLrzLTPage.__repr__r5) r6r7r8rNr9rrGr+rPrLr)r)r)r*risri)FrHloggingtypingrrrrrrrr r r r r pdfcolorrpdffontr pdfinterprrpdftypesrutilsrrrrrrrrrrrr r! getLoggerr6loggerr"r:r.rVr]rrrrrrrrrrTextLineElementrrrr1rr$r%r2r'r.r/rbrir)r)r)r*sf8                  D I%!! I "DD   {