From 9c235e7188effca26091e01928321f1a6da0b2fc Mon Sep 17 00:00:00 2001
From: Nikita Shevtsov <shevtsov@ispras.ru>
Date: Mon, 20 Nov 2023 16:03:25 +0300
Subject: [PATCH] add new info to tutorial

---
 .../dedoc_creating_dedoc_document.py          |  91 ++++++++++++--
 .../_static/table_merged_horizontal.png       | Bin 0 -> 11953 bytes
 .../tutorials/creating_document_classes.rst   | 115 ++++++++++++++----
 3 files changed, 167 insertions(+), 39 deletions(-)
 create mode 100644 docs/source/_static/table_merged_horizontal.png

diff --git a/docs/source/_static/code_examples/dedoc_creating_dedoc_document.py b/docs/source/_static/code_examples/dedoc_creating_dedoc_document.py
index 2e2cb1cd..ffc910a7 100644
--- a/docs/source/_static/code_examples/dedoc_creating_dedoc_document.py
+++ b/docs/source/_static/code_examples/dedoc_creating_dedoc_document.py
@@ -1,5 +1,9 @@
+import uuid
+
 from dedoc.data_structures import BoldAnnotation, CellWithMeta, HierarchyLevel, LineMetadata, LineWithMeta, \
-    LinkedTextAnnotation, Table, TableMetadata, AttachedFile, UnstructuredDocument
+    LinkedTextAnnotation, Table, TableMetadata, AttachedFile, UnstructuredDocument, TableAnnotation, AttachAnnotation
+from dedoc.structure_constructors import TreeConstructor
+
 
 text = "Simple text line"
 simple_line = LineWithMeta(text)
@@ -12,10 +16,7 @@
 )
 
 metadata = LineMetadata(page_id=0, line_id=1, tag_hierarchy_level=None, hierarchy_level=hierarchy_level, other_fields=None)
-annotations = [
-    LinkedTextAnnotation(0, 5, "Now it isn't so simple :)"),
-    BoldAnnotation(7, 10, "True")
-]
+annotations = [LinkedTextAnnotation(0, 5, "Now the line isn't so simple :)"), BoldAnnotation(7, 10, "True")]
 
 super_line = LineWithMeta(text, metadata=metadata, annotations=annotations)
 
@@ -29,22 +30,86 @@
     cells_row = []
     for cell_text in row:
         line_with_meta = LineWithMeta(cell_text, metadata=LineMetadata(page_id=0, line_id=None), annotations=[])
-        cell = CellWithMeta(lines=[line_with_meta])
+        cell = CellWithMeta(lines=[line_with_meta])  # CellWithMeta contains list of LineWithMeta
         cells_row.append(cell)
     cells_with_meta.append(cells_row)
 
-table_metadata = TableMetadata(page_id=0, uid="table 1")
+table_metadata = TableMetadata(page_id=0, uid="table")
 table = Table(cells=cells_with_meta, metadata=table_metadata)
 
-attached_file = AttachedFile(original_name="docx_example.png", tmp_file_path="test_dir/docx_example.png", need_content_analysis=True, uid='?')
+table_line_metadata = LineMetadata(
+        page_id=0,
+        line_id=None,
+        hierarchy_level=HierarchyLevel(
+            level_1=1,
+            level_2=0,
+            can_be_multiline=False,
+            line_type="raw_text"
+        ),
+    )
+table_line = LineWithMeta("Line with simple table", metadata=table_line_metadata, annotations=[TableAnnotation("table", 0, 21)])
 
-unstructured_document = UnstructuredDocument(tables=[table], lines=[super_line], attachments=[attached_file])
+table_cells = [["Last name First name Patronymic", "Last name First name Patronymic", "Last name First name Patronymic"],
+               ["Ivanov", "Ivan", "Ivanovich"],
+               ["Petrov", "Petr", "Petrovich"]]
 
-from dedoc.metadata_extractors import BaseMetadataExtractor
-metadata = BaseMetadataExtractor().extract_metadata(directory="./", filename="example.docx", converted_filename="example.doc", original_filename="example.docx")
-unstructured_document.metadata = metadata
+for row in table_cells:
+    cells_row = []
+    for cell_text in row:
+        line_with_meta = LineWithMeta(cell_text, metadata=LineMetadata(page_id=0, line_id=None), annotations=[])
+        cell = CellWithMeta([line_with_meta])  # CellWithMeta contains list of LineWithMeta
+        cells_row.append(cell)
+    cells_with_meta.append(cells_row)
+
+cells_with_meta[0][0].colspan = 3
+cells_with_meta[0][1].invisible = True
+cells_with_meta[0][2].invisible = True
+
+table_metadata = TableMetadata(page_id=0, uid="complicated_table")
+complicated_table = Table(cells=cells_with_meta, metadata=table_metadata)
+
+complicated_table_line_metadata = LineMetadata(
+        page_id=0,
+        line_id=None,
+        hierarchy_level=HierarchyLevel(
+            level_1=1,
+            level_2=0,
+            can_be_multiline=False,
+            line_type="raw_text"
+        ),
+    )
+complicated_table_line = LineWithMeta("complicated table line", metadata=table_line_metadata, annotations=[TableAnnotation("complicated_table", 0, 21)])
+
+attached_file = AttachedFile(original_name="docx_example.png", tmp_file_path="test_dir/docx_example.png", need_content_analysis=False, uid=str(uuid.uuid4()))
+
+attached_file_line_metadata = LineMetadata(
+        page_id=0,
+        line_id=None,
+        hierarchy_level=HierarchyLevel(
+            level_1=1,
+            level_2=0,
+            can_be_multiline=False,
+            line_type="raw_text"
+        ),
+    )
+attached_file_line = LineWithMeta("Line with attached file", metadata=attached_file_line_metadata, annotations=[AttachAnnotation("super table", 0, 21)])
+
+unstructured_document = UnstructuredDocument(
+    tables=[table, complicated_table],
+    lines=[super_line, table_line, complicated_table_line],
+    attachments=[attached_file]
+)
+
+unstructured_document.metadata = {
+            "file_name": "my_document.txt",
+            "temporary_file_name": "my_document.txt",
+            "file_type": "txt",
+            "size": 11111,  # in bytes
+            "access_time": 1696381364,
+            "created_time": 1696316594,
+            "modified_time": 1696381364
+        }
 
-from dedoc.structure_constructors import TreeConstructor
 structure_constructor = TreeConstructor()
 parsed_document = structure_constructor.structure_document(document=unstructured_document, structure_type="tree")
 
diff --git a/docs/source/_static/table_merged_horizontal.png b/docs/source/_static/table_merged_horizontal.png
new file mode 100644
index 0000000000000000000000000000000000000000..b22dde0be25eeb853094c12dc0f011db915feee7
GIT binary patch
literal 11953
zcmds-WmH_vwyp^s+=B*}1lQmmpwZwE+}&y1CAdrQ;1=8h!67(70yHj-H16(tNxpOT
z-s9W*+;M;1AE(CX)m_!K)>t*ys;c?ScZMs<OQInYBE!JIph-&sm0@52CeOda5#gVY
zV)J)M&tET`lqE%BDo2QSo+~fSMdU<aU}|Df?hWCd>qsCeO(z(bSI)mbFP6;6+@Bi_
zEmbv~RqQN5a;DCPCWg+2&uuU;UX;c*mLNrYOFL&LN<M8bN_R>=Rz^yX=c5nfpBfJr
z?|-bbv2*`dkH7l=uIu{fRyn<bfnl1H28yV<>m4j2>8Nfz3LlT|0|Ih3VSiFmGv`cw
zj9f2e!AAz{=!s!A^c<Vn%{+Y3a~cr5(5g3ZuewTJ4M^60xGQTkk)Wp(rA>~|qR*r4
zp&S|(_`%MO@X4`LS@QVol9R<e%V~2s&3wdr>rQYeP*mzY@K3?WRDR43`%_5Y;hFr^
z$ki1P^{45lghm|JpE8Gt@{aOPAu<$^_|wgVit*=PXW9h8z5dfHKDwHm>QAq4SpR!2
zhD;HEBz%g49PAzYXS~ItX!O>l#SU<qat9WFM(fFH<(S!_f*1bh|NmEWSV3tcsdW@|
zf=!?FgIAEmj?42QvQz4b5iTEQpc<c$lDxDv8>sXi`WY-7Y>B9^(qj`oeW<n_JUdlw
zE?jl)VTl$yWs$H1sEHQ31bs+n-d^N~A-y~mh~>ntv`LThJyW0r4N%K^=rJ^4KR=$2
zWRjZR1&BH_VV(CL3^gp;G67`%pLLC)dwBkH((p}%u54M05rwt*tAULyHP$ng=?$4J
zD)FYiLxt)NkLRy5e$iJ8LERWSD{I<8D5=p}^GU2(b5U+)%+==`BboY!dtdwvjh;r1
zEzj`1<Fpt@Kcc;DVV#~qdr_O9xX@gBzUz@8<b<aQN8J?sY^Uv$P8B>qgm^z(O?76f
zqT);HI9U2Rq6c1d>bDfq?IcT_D7r4=5^H$DXy&zX^^A&HcUP}|{`eBikm981)uODr
zJ7iLj)f_fVk}$?>PV_bEfxx`9W%T{}n$<L}&Iq08!7lw<NFJ9|uMc4|-0lX~^OMT!
z?>E+nO9MSv5SszvpmS=ZAQuP84^KkZ{kN#9#8tGNC8J++N1cQ$cE@pqNCMbAe<2YR
z@V4t-66*PIV+Hpy%(9uYQr=<oTq`K7s(E~;c#B7spJI@mXyeOQiJx`q92jcvh8ig?
znk9fCJLAA%2cH*^CC-4M-q_d><d^Md;3LsH578MZG}hpx$80~evshw7wLfZonLvlI
z+>ReSA|wZ%>?eyyxuU#Nf87G@vSW-v@0nP#16^1Ij>NN9-#CX0T)43kcMpQa*lZ`}
zPT>oBVu0tL+j!cXs73D~&_n#_i<ClKI^N*XuNR$8<F`#57b2l6U=61(7#zPy)2g-z
zH*6Hc)^7kOc*(q{j+5*2btW%my+<LFLZ1kSwyc1nbOi4f3^Cd9Us>LWK9(K9i;-9)
z*Bq^jI#BRec1m-bzBZRaLOD;_&M-SyT7g!KWlgn%kp+{JhN0v6vLnixu`mYyO#_gL
zA1d(^9iy(Pl*s;13&uGwLVoj*j?qODU+~^9lq7=q?|!@zFWrD2+|(OAZt|RZILbd%
zIh)kp-`o|z-c2}cUnJ5%1iZmM8mNOU_2<W-?ml7XbHr7iHY2G`SQ8=W{gTp!Ny$z~
z{c6#oIaU0SFK)ou1%8JC6VZ`|BZ|`6bDz)kFP?q|&uW(#qxGv9osU`0I@gbz>nDDS
zev0bI)^6$kZi!Nd5z)6A6nP5#{eH}^oiO~A-Kw$b0($$l!Re8*6QOL~lYD9hGLpQ5
zIg8c8mhT4NY;zy`WcXPzF5O*8rh<RKRT2xoecLZ%ap5{<t*R*XG^r(c?kTu`(a?Ta
zcs@zm<MS$TAh(TG@cZkoK2*s#CqA`k^!~)m>4I5jUtDV0d*T`G)+n}Ge4~+NK<)>f
z2nxWam|mvL5K<>Z#g-LxZcul(@nLvZ7{@sDM_4<<=@^|SPFEYV;Y6KO)Iy@MPhOAn
z0YelDYG=k9d|*3CAZ}zH^sL(}%-sQ}e)E3Bn5j_M=xTRr)_rB`6h31fJ5<4F8Mq$D
zL#t0149&GY37QVdKf5zrx!XFKoV_|eiZ#pei)7&Om7!-()Y?8cA1lW37iGFAYhpQY
z*sDxTA!C1?+xC^}yCy*&Dn@|QEqxSvHf$3E8A$&N=%CC25j8FFO^Q@2NiA$-p3iZ&
z7nZxjq#8t9ojxaVsX??NP|fqsZGk#z+|{gTrepr2@{hE;K|PEQ>PcB$S&a&qcV=%w
zdgxBZi=l|NcR?P=l}BH~I&w~CK}~SCQCU$Zf!PS?t%08e5e!Uka(QoV*b>CR)}HiV
zf}=Z+$;}I@2&0eb_?c>pcX6TDw;S&KTuQfZ2490#C*a)850D_mMWx>|;+l%a%vLN=
z?XFPTGp5MI`qBF%ebmXTR5&Dpy`bY?EyIW#kTwLhZ`bU~_1drye3NGeoL^Yy>%V_`
z8)`sy4No0Q_R?>PJ5~7YG+*ZPa#qa6hOw)pqH6DgM)`1nE}D;o87cpIy}mTU9tS1g
zdwDpHP5;`uE^laDwq%i(xqVs8;=5okLJIxpM`)ip0O%v<#Ac=+4so;&K8f$rEWs91
zI$6zVrSFTZp=_+wa*vrcwZ-gRm2gP<j`jNNAy{HTobgv({j$!;SoCUZYepAIcW{vm
z1@R+?$o_D0c6O~q(3E<(M3o&DHzS&|{;-eyHc4~1Y0_;M$UDnjUR&~|?{zIiIf>%I
z7A!(lVLYaBGOcYt(D2%5<co=tMOH4W9+1yEXtePm`KAyai)@DxM}`LMqV(lF#-uND
z^?_6?_u%GJql@|yW;R+e3&*eiFn+f}mm8VP*sus;nl=XZ9j6F1JeR2~qn^rBb#H$k
zODTM(><N@myCI;sI$_U=$0(9*=KB-qXoGj4qrMfeBTwbj6D2;IL*3ay8EI>Yp%?DP
z(rt<QEw>OhVAmRW#&!oxs{H1Shn(xYOOIE$xbNZN<03lY7&V*8&>(F`?E9f^rX=`k
zTX1@G;MBWt>;@Qq9TBbC13!anHHvk*=Ct~d<r369Clzn7SRF*mPi=NDO||dEh`%+=
z;O@jL*4q~mgWE0mQ51Si!)9n+fB)#%^QHP$bZYx!jksR&20QpV%;B!$sC@re6|ClY
zbg(Q_Ih$Od0A0NqI`#Rze3V*gXsYxYd!VJsSJe^|GB){=zP6;38ek5Vr%;Sb!ARgJ
z)JvbXu;TOISX=I2I>Rk$d0?hjc!_})96ETbICo4}cc!?qyJ2Gua30%Zv<}?cVvkGy
z9I&rUf|Cd1ag&7JlA3}&MRa)Tc<$@#u3Q*muffsGnJC=X>}#=B0$iFZo1S2iOQ|t;
zLyr1&zY_0szWLoi@+<>fr$;CMHLF_SU^=SN(V2g`O5g9S-r-6bVhMpuvgY|3b)K*k
zFRO`RtHm9>)8QE9=g0Est`2W(QH)Ton*;1Fn(q{Iz)i77Upig69e4J?a4jML503(P
zC5NiY_#*_iz80%{zAB*^EFW!oi!E=P8blEYPf6KB)~;K$#$U9gQEh;!vyswr#^u3H
z-x#9MDfJ^=5JN=`*0nHojNS&5kAEQg)x!kTeNN$1cKZXN@R9&P32t;H7qMbi>O_?v
zgSJvHnT)%a;lZl|T2!IY(C~BI<k}azzF-H;*^2RH@TjvngZ8{~TjN@=Bw(H!6}(8O
zwUj+sb8NkaWV-ZGr)}fXpey)cdF72a7Np3wxBjJ|4uojn>bJ{g6`Ni%x%P&I4!+98
zQ8JLGb>*(e99*(R>}@<yR3T=%+cMI(8%&|-%}cuI(|Q8wxVhf-p4v0_6=vcfEP=2k
z<70V9MR=fPLWow7SoH6A=H$0|w}@1=GPPAo*1HPc(cN(R<xzVydQQAG(Pird3BW(_
zA&`^5Z+9Vyv~6+)xqf3Gbs@46V23gkn46GyJMw?ZXSmA+lhhLgw*iPJPXQ=&hS_a|
zC}KF*j^1w9?!uv;&VF<~?8%^UC_~+A8ORLuzc8CZBL@7}*dE53TFDBvXKEJWthZ2U
z_|KX`8>kf35+gILf8-2zV)&Ss_qq)bop`24oUtF(%PpnBUZSIq)2MZVpp@UDWlF+M
z%E!^a6G6Vv8aDf0#1v^-RS=cjONCRjT=k=jq#XB#yu|77uanb`(<|cddKzCx7Xj|h
zEW0wQPR9iAB-{TY$DQS6l@`w}vrgiqd7c-gNr!QIGs?k8sf29sL=j@ut7GCg$w_%{
zyx=@Pw!2DeVvNQv``1kj%<|MC47R7n{Ad?I38iPwX|hMg{z_>y43869-jN#_Ov}I|
zav7mFy%+dM8%*Ln9x(an<j`9D^6lFsyIv2Tdh+*KC!I>uUZiW>1<*QG|KWY55SFh`
zxGbZ5sykW*cUvNtg<bMAOp<6%P5nx$P0wa>SYl&DY@8@79`lr9toNzG&92Z6?@j1B
zkqwkdl5wdY@G3|Ya<AwL$_of61&pMgDk;9nadsA6^t!p6@k0WL5>nhnqRqx91Vn4c
zbk~^kh#sltJ~7K!6pCt7L@237O0vCuHUa(BEUrli*^j;kVY=h{WiyIzKG+)Y^IU{>
zyfS&EBK+rRr^4P`ZP-yeUe!l59?Ze7WR$4i-cRv*I?%x1Utcs_OP>flaxuI*NY@B`
zcs}hr*N?zK?DmQb3SX(dc>Q&*bj4he7NOMBAET^o30ON`m$`WiNtmXl*WJ1|{&&h7
zJCSN5u)=T1Iv)-PcWo1pX_H`0m21>8E0w((L?Zd93&^g&4#u;pjQH*=aOOTvkgPZ<
zAmk8(ojQg<EssytW;^sYBgLUPVJV9GS)NbD(2WUH*FJ0>?_w(je0-e2tKaNn6|1Ea
zr`vr(l8#A6X5VHD30kB%&9zyN=zI2@NM|EIz|)ER5{!t|)AtWb54Az%IR(ENUIO9F
zZPs^@)m)8f=49cw5^2U-(GqcYV*l*w45DgG_9K0r)-=^PvZa=2u|7|rH6cwr>m}`J
zitQNd8#IIe22w;N<$}rchFGsD+xWIP*5T)ANP(OC!5de<Z?Fpr%SByBuBej7;DWhX
z5XZu)FR~v5#ai97dp=2q!|wEDoC_wBWF^IuH~c#%zayo11W1YtEZZ9hN>jE%e{n2^
zW&_&J$I47dw1S#VKiSS;i+#c?Q)7)5`WKRw8*APY)Q7VlV?^9J%jr%V6#9=%z2ei<
zqgvUKG~U_J*W=u-euuKiX!p0(9KD07;<$rbSPJhQ2w|b1{HdAc=r@HH0hG$OzV_a9
zzh`+%Bo4QAd=aX2?X*Jdf&RxK&BwztLFVDGfZ;cWf8?ixg+igz+RMJ~KH|uzl@BMQ
z1DT4g?ZR3dt|G=&nB@*$0PyH`?v!PLIl%QR_O(06tC4SB><B^#*5@@`8P{~51K%Rg
zUjkS;I$reS!RLr9Zf-x=l_q|9N>(lkl`pK|OKcGZWU9fn-ROvx-09vuVJGn88v8uq
zS?rG!=uud&TuN|c#b*^vXjFY@o^nj2&#WVkZY53=Max_g2x@ilM`RBrb*eE_vZt;8
z@xl)uVKROx!KZtAo70kH+^^|Z4c>QBSF4?iLes_5LE1u(W7MQ`zKyDiI#ik5@j<bg
zWfO#n@L0CSHU{E@3y>X2bX5kt(N1}a;eia(JO)7oOsKI2!5EB7`c+>`NyGktHnV;%
zSfuRpZQuBH&fWM6!OhCfyd`8Cziu*SbeDoFlh}|7FT?R>XGkT((R|7)Q_gy6%lvNt
z+!r98ddwGSyLlc*ZiYLh%iee!8JKmn=x2J$#{E5Z$FB8(PG|3BKiC9pKiPC>Kc12q
zwr`x_j!F6CeqQpzo!??D*4L&*qi%$k%qa?l*twQ%(0&|~9s)}Zl~{ay&*THYDg4^-
zwV(Lu-Y9Tx1(z{~B=(ZG4C_MH*fQNV#Lk{NZ_b}}5A}B}<$G9{z+|5HQ+$yk?eLoB
zVC*9z_9=S;h?WY&MeBA=)4o3)PPXYY15I3_H+)km%~-QBf;<0%m2<$jjdylxP;DU^
zxKc&Z9kK@v8I-;GDa#iW)yikF^TgOna^#n3GbkLIFBwBsCIz5rpn-1f!EJ@UGZRIs
z%#ad2P17mPvWSi$X+2q}yeAEpHdmUjCgF2AZ``8d{wQ$Vo>-shB<Zm3ju*ip5Aqtv
zXxAJ0h%j=kd)O`~BN)7YgdOi(kAt@0BHC|d>kPl1B`<M`p>F7W{(b@7H*R$Eo<z3s
zep$$ZbD!YW&jpt3{CW|kG^;B+&hgqyhhtK*Oxp+H_a1C=RVD9!O*-)$c{)tnAzt$9
z#9w=tU%hti=zKTzg5!BEP->S9xO6=T<A?#je?IAV(YTXCY8AcB!;WP8JyY%io`X;y
zW9KojYE#zoPhKQF8nFKkN<Cl+n8yEJo5HL(^u#XXQu)99scp|_#0kC>5S0L$MZp18
zV@~|ThOHbKS2#`nM7Mzda}X?V{--li_})|GFS;4S&*@>?U=;?GZ;&28<IUlxRmFD5
zdthvd&YMRsYO}|yxck&FQthVj>#O0kYkM?#bXoO!L!(&v?KW~HyQauhfQHb${dGUh
z4(m@j;Q4k*FkBiIKDH67{gTZ75Z9nB@Lc1a9#C{xU|73J!HztyKk38>)@bur_n^<5
z;|DZVPuwkq?9RDz*rrcmR)DN;QqOm<3ZPxgtA^U(jVHqmG67YICLaUJ*U6JHhyswx
z6jTLu?1y{$N|f9my1T~#*p+0_-BHEl0t<tHO6Eh0on?e4jqFcGdk}vLxEiYb8k1}s
zL}8LpJq3@hU)}hNp}pjE3R#k7!}GBlQ~}_cLVlga)|CZb*oo=F?q=b#mk22s-El}|
z4tcxC6i#cZz_v#U^prt4ptF7wu}+Mk7`1`R6iYS^I+u~v<lgnz{RX>kKe+LA7_ui`
zg@6upb?Mck&*P_y*WCG^aL|p|-gzQjnifJQ9NLZ(i--igTZ}biU-{yqhi%8JW-hN<
zt-~vu#lPZz1^#eJ!Em4wwReFduqd#y;6q^Kn@()-bc%&|IQ#0b((d^5AS%pqFOHdP
zzf{fp-G7HrG(AcSXaPh|uRjSlBs<1@Wf~KrC5*&o9>g}K#E>S7?Lj%P%WlDpKX^8x
z>_Jf1xr@8~VJUrZ44W|;9<yHL5z4L%(yT}M7gnrnefv)-{lk4-ynd27%vc{nQ9JJQ
z9Ss--g9RP3$;}*TJ>?X{YpdA^NY|szKjvcznHIT_*6y{ivauqraqCD#3gxqT+#3kQ
z0^H3F6i_uhB#<wxcyPZOFM2uEl<p!M{_5)0xbu0jTe{9g*<3X#A1tMRMY>tZ7ZqWj
zjrH0Z&*)_SNQvC%Ms`cxu&Za(p{;j_fZdB?^Ij0Mxu=leoq-U2SLgp2j}S<DewEdu
zmyN_Z9RYNo{|XdPQ`OBl7jU6;r$R+Q^7_GGsEQ&MdZqD9&-<6r%l<=5w+83yCj|Eb
zpq0!|+q?#EM+nOBvlVZtT&wh=qY-@T<+*e_=C|?3{01my5!xv>hyz!uZsE_0dbEnP
zQ$d!t9HLvMD+BEf1G-^9>w##>ZBv|K=MDUb(bVr5`An+?gxc!Ls#Wg7kPu|bYs22f
zbCtD?&AMO@jihd2<y}kEuxcG@zJ9lf;uo#vtog;#mL&>5encksEd15UR{1DaUnC`0
zW2K^3Otlc#k4wl_ko<&l)ZAadh-d$O2I>2iMNN^$m%Vdk@RU50H#IRP_YAFVvPD%5
z_F{5oe6lFeP5EHVR0`kTZ46PagYxFYvZk&qQ}Ip~6BkM@H*8gf<+_-RWX(g1#~o1y
z1jMr~>s!?6_KeF~dl`BcfYf$_6jWP_5fT2$@#J}-dCeysagX)eIp8pV@uKqBvJ-9Z
z8OVE_h-)eH+{v4;;%leVqXs3NsQ*G}<{e3WBdKhLtO-lt5L=75xGPAE*!S5KC2hSH
z){da7Ym)(SjxKYE@7{X1inz9=&-{-hM)68V?w;PWlnS@9q!!F6bk6efl2x|G5Lvh?
z^;^pKZt&V`7+P0+uVakYl2x~xgQ$ilJb7e_W;D_9lXZk+qv~)9Oj1OPFlm1dRdJ}6
zBWHDrVeL|obt!Z8s2m97G<we6&L##2>C&-&NAAuUk)P&bf0#h@o5P66ZttBDCd@hY
zg!2J7^VD*ZB2+L0tr0blf1H7X8fJ}`VS)j<r(Fy_fE@l>+>`0Q7K9I_=(=6q251O;
z+b+o#{(q-m_!u=0N_Pv|=a*Q;Vk@^}YKYUl117Iu%EdRbEQKhwTZA<v#PjEmWf>0d
zQW60ib%(ikjj2eQNG)+jR>9RGnM&~_+b7P~L&qxjLH0i$YrdS6yu)VRe}bohMy?3i
z7XJJwkAZ(Baq3gsvd#=Naf%fYp>7-a>VpZMB28OZjn|-rYdn{m{cIL{h~*a?2?2OW
zv$<C!@DmC^;g*F3Bjp9RR?^>{;W%~rcxgq$vkvC-5;Cvk;dm!8A0De}r<YHHoUCCB
zclatdr6_dfaC|76yM&yGR-bJnr#)`4EC}nzq)$QV36p3sPXnuaz+S^7q(t9@Dz#h(
zXL-4Zb)t^XOTI}#x5?@3Xzx$c>U~mV&Y!!LVGp_j<pgsQgPhg#2y+R4^#0)Hj{6L-
zFi4yL0GM@#LzH%4uSy4AOxoF5<Z(c6r-*lyTvk;~;Y(JAK1tJ8>*$s-U)h!ZHD<}n
zDx0XWutN|}1IY;javU{WqjP^2B8IKafq&Oa*=T}5v|b7ghiyvaAsMC5ZLn>1$>!?W
z$_N!|FI1FXAeZ*=VxOW!kSE4~O09jx-x7vO<g|42+QbU`2XK_UmvsY<9xHPA3J76Q
zrYmuwiv!&dG8RZ<0dSVeu*vBCSy05JAKt-8Dp+i0fCw?Qm9bvY#H3_$FL9O<$a8>p
zgZW4A6$5vio@CLKs3E<m$kaN_*LzqTq_@PGyIQ2nOhO+E6CKgfaPkPm#qVoJft+*3
zpC1{(v~OTpcpjn-_KdI^Zg_U#F2q29J=N}5{k4T`9QQ~2TFkk2r1MRQ{1;cPM^XpT
zTui$iiM}w)sj)#x>>CD8kE*G=!qMnL(`_1q5)@(Xok+s=`k}Lpzj*Xm5FZBB!@@|3
zO{s!_tuRH|4aT}Q#9U-mp*g;t)xLMA*C%`SCkd+__qJ+p=DBNt06Joj!aF5f=c47+
zyJVvGXh&R3iv5W(faP^nzA11}#=~8_RcUMy|FuT?AwjmU92!fuSv7lB2UP6GPg}KK
zxZqPk%2FiK>@~IvFN!>M9~}*TY&IG~%0_+6j1B2QiAsxdPU!llwnmdUQIkU@@AF&O
zHyR4ic7zy2Pfk?1vj*eaW7FC6w9wU7b}5Z_s45LOU02_s?lmz247>U2K$giL)p+d2
z@;Qbe;g?BOm{@{}7hd8y>JKOy{Os^Fgz)*xzgV(gMe7iWS@VShyjY|+(vX@Yg}e4p
zUeQN;N&vt2ANWNc>D!T;B8XH`mRkzdE0D#Qda=i=XN0mNePFfO#lyL64}szibds!2
z*#BQCs1(V{ROIXQf=I=bNR`@8CNp_=2M|H)AhOctB0j*2=jVYkP2*+1FpZn&$?c;B
zcor_YO|-i201YUdAF5XL$&^m!EclV$;v#AEqvHGg3DLr;#>8nl3#`@6wwD(%;oF-B
zsQg=-h#ARY)ZnMikcRsTuUQj$7IcYJ_1DNltzb6;4ApWhtVJ@b?SskjtHx7qh1GLP
z_)w4B@@mv2AA~$gVp{0@Q+6t(gkbNk;k|K6-!ZZ#pn*H@B<g%|#<HAEoQ_GMXgtq~
zXY75^M`?!u8tzwupC>(MOA3!_+IKJ18t#}X)$GW7QO`Qa=qo18JZuWbwsW3)=Cy+9
z5F=MNp>HqHW%oDH%=yhZFfu}VsKM8xAq^|~oV<;4-k<O3@s+sT;UbOu^L<9FfPxbQ
zF2w1%3+dq&T;#*=+M1hCu;!BgMi__Oe<KWG)e1sK1uEVf1RYMku6$dB>wzL6EHK%X
zf$^uF#Mt2iv7bN70+Sy{yB1DPc+>ogzm`ZheM%%=adv#?+h`S(FRAT(oS+bVpO5S(
zItV>-04tJJ@VWXoOxTB*yeYjPb+<c)sUfGzBMhza4;y47#f;aKi-Ch2eE8#U|3@`>
zveEkIA7?$xx8kdwSf7YsSaF{rBYB=<6avuXcl9q<KI{3qzX^fz-(=@vymz`XJ-1JG
z5`!)Y5K&-ko)&Mhq=TnGa^KrE{QPs`LS!vT{JNcSqBekl0n4ZukW>AF1;QlZd9U2_
z_6e2>N3paH<ZKzp+~!5q5eW7#<|=j-pp}=rALn_aod77Vivd|qz_ocH<d_9L`_g~Y
zU6=L1eUG3$zn4la6V>OVZrRzz)N47Uu{cyCHoge#_}>1FEN(_y>e@A%`q<zfwHWjF
z8-N15q1ulhkX)lLg0(wr-n?F?rYz-+GOxbLOw%|PA#R5>Ve8xK2KveE$Bqcu9;aX^
z7djr0Chm7>=ySgumWS<rq4djGp!W6Au#P{*YJ1P$wONq;z_VxF*EPcgQ%mOPsOWhi
zAfL?}Sq@8jLRl(jwM7@XxvMdSo-ZsMR<N#`(VQAik}ywUzJL?;a2=)k?PJRO$0<gg
zri?(!KA}9q0YY?%wX;BJDS5%*$TWV|5?%Cqxz~}gqpHG${}5w;-;>;%0>A$jc&RQX
z#Z3-VgIswRSWWl)8yi5NKQi!EE;WKoqquFyTsa3L%H)vkGrS0p&cv0Z=kzCHZJViw
zb>Fp@%`(_E`Y%T2V_nyALv#ekzB;`zvk{7Zbyl}(U$g&R{lw0QB1+xWU}Qz=oQ#Jw
zJBvfo2B2&tN$EMDm$Af)P2qb8x{B<(9h&a2`A-;=-yu-*Ph46Yvik=vIRdZ*5qGHt
zSxtf5v$&*v?ML*N5k$=UeXHB=1r842)nmD~iFpC(3fg(kCQgQaS6>!4%EeN59qwHD
z`AER?sG`v=f@~xyVLY%jG0V&T()aN2I8x8vgA_|}aMEJj7+J!(RtyUDX?%@-m`Eh`
zx~E}Cue<DsI&aJ!j(-zX!sO`_!PAcP(1@Ox45NeuOXt!AKO1`w#^IrFn4aI<=`jB$
z{P;KqGg$})w%Qu-d2>u7Dx%!Iz;I|()q1Q+A`o;Y<jSm26~r?&SQDl?-kKhFOs{``
zmt(;w-$<)a_sT>1+I=0xVrvH;WBHH(FCj?4gAi+OX^p)&^v2sE7}hEi5Bbb^g43Im
z8I!{4^UA9mE+aKY(YjF#ca~N0sQE8c7g2bg{DK_xmUh8}T5Y7BZGz{?D%BqxtQ3fa
z!hNvg#a45?Epb9Z>wB@aO8Bdex_Uauqn1IhnQo){)=vUY1kj&JLXgd1<2AvQJguyL
z=+$QBx0f@HeNmhmjqXhCEX52BnBYL8_IgzcpDGuv4;*n_o$3Vm1OdF#%wgXA@C<n3
z-CLnom?x{WHow@Uq3<#H+aw|=HbLaQ@J<5NGu?#f@`ef!nlK;sS7NJ)?v|6Fpn9LK
zZ#tBolWA2<AElP9Nf@CkcF(IDm*sCww}621LFGeDCo#r;#qX(eNq(51(fXSArNdwf
ztCc(Lyby(4i?$s(Zm4|lj>grJPI>7UVuOFcTEWIsBT<>Ueb1-tjey~QRS`q#AR@==
zK7ltXMimpkWK#+}MA=x@q|Zs%iA6?sG&uQ)Ra~_`h){aI3?nS?EVIVkv3J8Hl=Ihc
z9wcwrh%9K*4N4f)JlFFOAt=NXyVW_#SQ2s8c9roA83ySLd&x8GzVxN#hWtaJ^Lgu-
zAf6i*>x_q7n6TYUD~B^wkvx-imoa9k6~Wfxw81qH-Mw0I$(MUc%5;fC)U#8l8=N^W
zDa)hSKK3u@1<@g*Y>u`HWfp68ZwZvJfW-@s0L8u8bai+|zqMqv0r!EU+S~pDef;WZ
zQ+5-?svuBQK?N&(KQ|9rV`s~0m*60)RZvAkIH4TZq)3$4?aix;uy_x|kaSu5^hc1)
z$9fS-?|XHYuo=NDPY;E76|#l~?ry+8U+ewk3<th(DK&yJPZ-V!av^B9SxPl1aCaSk
zT&E}{qSDn0$`0-LMWB+*^{fZZO7A)KeG^LFu=eh0e3<4PL|4GI<<tT{R;<f3{X}y(
zxi=jW9J-guJPwb*-&*H3X(eC?*eea*3e%m2&(q2aN#guCxkP31Rb`gd5)mFP>V4P`
zmS*`Ubon-XgHCdFTEngl!IpR<8+hSZVq!TN(a$%o92g^Vd%5Q(rLA%NDor8AgHpoV
z>;0^z6o%@^vK3vd#kC0#TCi~Ramqu^^^UgpL%#nEgAY~?e_+K4g5ey78Ba{KwsP*?
zs~A>=qzyWt^HZw4ZASq`AS)~7%g1(Iwg-NibG>SqK%GKNi{>c778L`VO@!Et{AmsH
z#h6Xre9JcF3|T=6mJNAT5L$LXyf1hD3oOCUvOM1nUxlC&hJ|7JUAtz+9A$VKcLs=n
zJ}zZL-pin43YosK3kkH^fQK(`)e~8UAsMlW?iaCk^#3QfDJR}={tGZMIAU!-2W+I;
zS(N^OnZP<6Es;>LD=D&0TZXMIIc0iT8(bQ#n35fO<%2Rj!jJaUnoIsS_TD76jOMth
zAi6zdS!D{Gi7IwjSlD%K^{B;U6#gxl7iZ`=m*^nHA&g;Ed=T3LVqqk|07HWq8g~P=
z;+M*`(wp<5e}ytZCZA^^a4z4gM<056odJ|N9}D|NxxT5>&0>mvyvGl){_&x&Z)QP5
zftN#9bQ%{MwcmKG%E^N<&Q7=ANV}3GBXgf|sF59RpM<!;o)1fK{^ekVW4zM=ej@L6
z?orhDYi18By@odkS2Vw|ibqu*y)=zTYT(5!&^g#=ldI?15wMy641afsb-l-w@QhsM
zTO9>9i0U(5;X+y#9R6U99X<}qZzc2Tk7ZiMQu|!qH5csV+D!xrUD+am=`ATZ)f3X4
z>Ibn+{gH~w%E~)xQLV-(>o<q-{}HPgT$;YIiBdIpLm{p`WVG++zcNX=?S~(a^~Hk4
zeSjFTr5Bhzc6YGa!rGxpku((dY4EsS>A0@dA%RUidaAs5QVpvD#M+qTqf$W5Y~+g2
z9-{?(?UT9YBn<&N;!%h!x=qIO-t&HK4AI$*C-mS1K{2i#Nqd|G%r1ZvZJ$O=TexAB
zgV{^FsyGJZoHzdRRJO9SfIPB_@P_P0+^&T)SQ$x4EkyY&t;8+=QX(DC#Xv^b-+tJu
zWiZXL+Ru+(_99yXf$U6=@hxF6a0RX?yT(VCag~cqvMt9K(as<<GmMlniN;%-SL{A6
zia!(pmnie;+c#xZ8e&W(Ih6-Uhk#<La$;^d-+;(>CjcG41x`R1e#Ud2qf1!yLe$zt
zz}Qj!z$ZdIp<f2(0-w4=ODpXiL~@_g_pAc?Fa2dRJ17vZ279+AB%y0!bEn^RW<ZWt
zG#WJzb*7|FSAW-pj_U9Su|%puaxEWSWy#u6z1YtbGu+^jK7<(VCW~ITa}l#l$D3Cy
zqhPh|u650)h4F$hJ0Aj87!@e<LQBwLLxZ{*uAF4aF4)jbwIzAjb@;;<v&%a!Cr+7d
zpLL{e5!e`=q5hjFSqR(n@S;`PARWZ$UsjMzpaxx7Iz!K4m|-1=5HUo|&i{5B7mRYj
zQr<h9ifusv!4PqiLf?fKrz`+-dJ8!TJf?h(&I@q5zU`*H=LN+GAYe_s{QiwxW$z$o
zdN)U1JG8F-Z76AKiy4+DL;V_>A3o2dLr|NKSluBf(%$&c6^A9*w?pwI4{>*wKYZ~*
zEBBHd=hG+5x6Ym<s#+eidomdTcocR^9~6&L!|awus-5g&IxNF2UOkZ6xDS!x`VE4S
zSkTb})io&wTknhvFc6wD!w$(f&t0u7Ts$GX#KXN`#M^RbAK~wHi52_7q?c&g`7;GB
zQcvPbfJ)og7h-@FkPQVa;iUN4(oq*7dVxm1iU<RSmmDyb_P1(Oaei0PXWK3E-4Wo#
zL+~u&p!3wT!l3PPNglbWF!^b#()AK%$}l{IXCP0Fq7``P6GyRDy~9cN$Uo{JEzTgh
z_q-W|uQG7hcdr}Oj4n1O{7_T&y16Ya%nZ;~K4Je|2pP>;D+#B8pf-#Y7x_#v^7KJ6
zOEi}?)|^hbSwZ%O$L$<3#vE5<Z9RnM!5g}vbFn%*zJC;->Wo*2wvekLi2;GcyC8%>
zCZyyaU#LgN99t~B@Rhq(zX(4}L3oAo=Igx^kLNniw+If9qt`3ncD-UkbeZ@EAIi@o
z$$5q?^4r@=nZNhJ&lS%BSl*7SaA_XxdCUVo7oba5CtwKe-^i-}nXC>-Rqc*hpA&EY
zMpm&0|1D+I_x*c5uT*qP$X`jm+cQ?XoF<&5|AkPca_)na|BJJE|Gy<IBdt8asCfi4
U%5Ok1e<zPii^~HmMGXS}53$l-bpQYW

literal 0
HcmV?d00001

diff --git a/docs/source/tutorials/creating_document_classes.rst b/docs/source/tutorials/creating_document_classes.rst
index 04a0c588..11b95ba4 100644
--- a/docs/source/tutorials/creating_document_classes.rst
+++ b/docs/source/tutorials/creating_document_classes.rst
@@ -4,11 +4,11 @@ Creating Dedoc Document from basic data structures in code
 Let's dig inside Dedoc data structures and build Dedoc document from scratch. During this tutorial you will learn:
 
 * How to use data structures of Dedoc to store text, structure, tables, annotations, metadata, attachments
-* What is Dedoc unified output representation of document
+* What is inside the Dedoc unified output representation of document
 * How document structure is defined
 
 Raw document content is stored in :class:`~dedoc.data_structures.UnstructuredDocument`. This is de facto
-a container with lists of data strucutres objects:
+a container with lists of data structures objects:
 
 * list of :class:`~dedoc.data_structures.Table`
 * list of text lines :class:`~dedoc.data_structures.LineWithMeta`
@@ -21,17 +21,19 @@ Order of data structures in lists doesn't matter. All document hierarchy and str
 LineWithMeta
 ------------
 
-Basic block of Dedoc document is :class:`~dedoc.data_structures.LineWithMeta`:
+Basic block of Dedoc document is :class:`~dedoc.data_structures.LineWithMeta` (line with metadata):
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 4-5
+    :lines: 8-9
 
-To specify hierarchy you should use :class:`~dedoc.data_structures.HierarchyLevel` class:
+Each document contains a hierarchy of its elements. For example, a header line should be on level higher than common
+paragraph lines. Hierarchy level is produced by :ref:`dedoc_structure_extractors` and may vary depending on the type
+of document. To specify hierarchy in our handmade document use :class:`~dedoc.data_structures.HierarchyLevel` class:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 7-11
+    :lines: 11-16
 
 Hierarchy level compares by tuple (``level_1``, ``level_2``): lesser values are closer to the root of the tree.
 ``level_1`` is primary hierarchy dimension that defines type of line:
@@ -55,20 +57,33 @@ Define metadata with :class:`~dedoc.data_structures.LineMetadata`:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 14
+    :lines: 18
 
 Also there is an option to add some :ref:`annotations`:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 15-18
+    :lines: 19
 
 Now you can create new :class:`~dedoc.data_structures.LineMetadata` with hierarchy level, metadata and annotations:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 20
-
+    :lines: 21
+
+A few words about ``tag_heirarchy_level`` parameter: some readers extracts information about hierarchy
+directly from tags in document. Dedoc store this information as :class:`~dedoc.data_structures.HierarchyLevel` object
+at ``tag_heirarchy_level`` property of :class:`~dedoc.data_structures.LineMetadata`. List of readers that
+create ``tag_hierarchy_level``:
+
+* :class:`~dedoc.readers.DocxReader`
+* :class:`~dedoc.readers.EmailReader`
+* :class:`~dedoc.readers.HtmlReader`
+* :class:`~dedoc.readers.JsonReader`
+* :class:`~dedoc.readers.PdfImageReader`
+* :class:`~dedoc.readers.PdfImageReader`
+* :class:`~dedoc.readers.PdfTabbyReader`
+* :class:`~dedoc.readers.RawTextReader`
 
 Table
 -----
@@ -77,65 +92,113 @@ Imagine you have table like this:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 22-25
+    :lines: 23-26
 
 Main block of tables is :class:`~dedoc.data_structures.CellWithMeta`. To create table, you should
 make list of lists of :class:`~dedoc.data_structures.CellWithMeta`.
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 27-34
+    :lines: 28-35
 
 Table also has some metadata, let's assume that our table is on the first page.
 Use :class:`~dedoc.data_structures.TableMetadata`:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 36
+    :lines: 37
 
 Finally, create :class:`~dedoc.data_structures.Table`:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 36
+    :lines: 38
+
+To place table to the specific place in hierarchy create :class:`~dedoc.data_structures.LineWithMeta`
+with :class:`~dedoc.data_structures.TableAnnotation`:
+
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 40-50
+
+Let's try to construct more complicated table such this one:
+
+.. image:: ../_static/table_merged_horizontal.png
+        :width: 700px
+
+First steps is almost the same as for previous table:
+
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 52-62
+
+Then change ``colspan`` parameter of the first cell of first row to 3 like in HTML format.
+Set ``invisible`` to `True` on the other two cells of the row:
+
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 64-66
+
+Table is well done!
+
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 58-69
+
+Add to :class:`~dedoc.data_structures.LineWithMeta`:
+
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 71-81
 
 AttachedFile
 ------------
 
-Also we can attach some files: TODO что такое uid
+Also we can attach some files:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 39
+    :lines: 83
 
-Unstructured Document
----------------------
+Following the example of tables:
 
-Now we are ready to create :class:`~dedoc.data_structures.UnstructuredDocument` object:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 41
+    :lines: 85-95
+
+
+Unstructured Document
+---------------------
 
-There is an option to add file metadata to document:
+Now we are ready to create :class:`~dedoc.data_structures.UnstructuredDocument` object:
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 43-45
+    :lines: 97-101
+
 
 Parsed Document
 ---------------
 
 There are several ways how the structure of document can be represented. In this tutorial
 we will utilize :class:`~dedoc.structure_constructors.TreeConstructor` that
-returns document tree from unstrucutred document:
-
+returns document tree from unstructured document. However, we should add some file
+metadata to create tree representation. File metadata is usually extracted by Dedoc but because we are
+building document from scratch we have to add it by ourselves.
 
 .. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
     :language: python
-    :lines: 47-51
+    :lines: 103-111
 
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 113-114
 
-Great job! You just created your first document in Dedoc format from scratch!
+To get the tree as a dict:
 
+.. literalinclude:: ../_static/code_examples/dedoc_creating_dedoc_document.py
+    :language: python
+    :lines: 116
 
+Great job! You just created from scratch your first document in Dedoc format!