Welcome at TectoMT Tutorial. This tutorial should take about 2 hours.

TectoMT is a highly modular NLP (Natural Language Processing) software system implemented in Perl programming language under Linux. It is primarily aimed at Machine Translation, making use of the ideas and technology created during the Prague Dependency Treebank project. At the same time, it is also hoped to significantly facilitate and accelerate development of software solutions of many other NLP tasks, especially due to re-usability of the numerous integrated processing modules (called blocks), which are equipped with uniform object-oriented interfaces.

TODO popsat instalaci

Before running any experiments with TectoMT, you must set up your environment by running

source config/init_devel_environ.sh

TODO obrazek

TODO malicko o TrEdu a obrazek

In TectoMT, there is the following hierarchy of processing units (software components that process data):

• The basic units are blocks. They serve for some very limited, well defined, and often linguistically interpretable tasks (e.g., tokenization, tagging, parsing). Blocks are not parametrizable. Technically, blocks are Perl classes inherited from TectoMT::Block. The blocks repository is in libs/blocks/.
• To solve a more complex task, selected blocks can be chained into a block sequence, called also a scenario. Technically, scenarios are instances of TectoMT::Scenario class, but in some situations (e.g. on the command line) it is sufficient to specify the scenario simply by listing block names separated with spaces.
• The highest unit is called application. Applications correspond to end-to-end tasks, be they real end-user applications (such as machine translation), or 'only' NLP-related experiments. Technically, applications are often implemented as Makefiles, which only glue the components existing in TectoMT. Some demo applications can be found in applications

This tutorial itself has its blocks in libs/blocks/Tutorial and the application in applications/tutorial.

TectoMT blocks repository is saved in libs/blocks/. In correspondence with …, the blocks are located in directories describing their purpose.

Thus, the set of TectoMT layers is Cartesian product {S,T} x {English,Czech} x {W,M,P,A,T}, in which:

• {S,T} distinguishes whether the data was created by analysis or transfer/synthesis (mnemonics: S and T correspond to (S)ource and (T)arget in MT perspective).
• {English,Czech…} represents the language in question
• {W,M,P,A,T…} represents the layer of description in terms of PDT 2.0 (W - word layer, M - morphological layer, A - analytical layer, T - tectogrammatical layer) or extensions (P - phrase-structure layer).

Example: Block adding Czech morphological tags (pos, case, gender, etc.) can be found in libs/blocks/SCzechW_toSCzechM/Simple_tagger.pm.

There are also other directories for other purpose blocks, for example blocks which only print out some information go to libs/Print. Our tutorial blocks are in libs/Tutorial.

Once you have TectoMT installed on your machine, you can find this tutorial in applications/tutorial/. After you cd in to this directory, you can see our plain text sample data in sample.txt.

Most applications are defined in Makefiles, which describe sequence of blocks to be applied on our data. In our particular Makefile, four blocks are going to be applied on our sample text: sentence segmentation, tokenization, tagging and lemmatization. Since we have our input text in plain text format, the file is going to be converted into tmt format beforehand (the in section).

We can run the application:

make all

Our plain text data sample.txt have been transformed into tmt, internal TectoMT format, and saved into sample.tmt. Then, all four blocks have been loaded and our data has been processed. We can now examine sample.tmt using a regular text editor. We'll now stop and describe data structure in TectoMT.

• One physical file corresponds to one document.
• A document consists of a sequence of bundles (<bundle>), mirroring a sequence of natural language sentences originating from the text. So, for one sentence we have one <bundle>.
• Each bundle contains tree shaped sentence representations on various linguistic layers. In our example sample.tmt we have morphological tree (SEnglishM) in each bundle. Later on, also an analytical layer (SEnglishA) will appear in each bundle as we proceed with our analysis.
• Trees are formed by nodes and edges. Attributes can be attached only to nodes. Edge's attributes must be equivalently stored as the lower node's attributes. Tree's attributes must be stored as attributes of the root node.

We'll now add syntax analysis to our scenario by adding four more blocks. Instead of

analyze:
        brunblocks -S -o \
                SEnglishW_to_SEnglishM::Sentence_segmentation_simple \
                SEnglishW_to_SEnglishM::Penn_style_tokenization \
                SEnglishW_to_SEnglishM::TagTnT \
                SEnglishW_to_SEnglishM::Lemmatize_mtree -- sample.tmt

we'll have:

analyze:
        brunblocks -S -o \
                SEnglishW_to_SEnglishM::Sentence_segmentation_simple \
                SEnglishW_to_SEnglishM::Penn_style_tokenization \
                SEnglishW_to_SEnglishM::TagTnT \
                SEnglishW_to_SEnglishM::Lemmatize_mtree  \
                SEnglishM_to_SEnglishA::McD_parser_local \
                SEnglishM_to_SEnglishA::Fix_McD_Tree \
                SEnglishM_to_SEnglishA::Fill_afun_after_McD -- sample.tmt

Note: Makefiles use tabulators to mark command lines. Make sure your lines start with tabulator (or two tabulators) and not, for example, with 4 spaces.

After running

make all

we can examine our sample.tmt again. Really, an analytical layer SEnglishA describing a dependency tree with analytical functions (<afun>) has been added to each bundle.

You can view the trees in sample.tmt with TrEd by typing

tmttred sample.tmt

The linguistic structures in TectoMT are represented using the following object-oriented interface/types:

• document - TectoMT::Document
• bundle - TectoMT::Bundle
• node - TectoMT::Node

You can get TectoMT automatically execute your block code on each document or bundle by defining the main block entry point:

• sub process_document - run this procedure on each document
• sub process_bundle - run this procedure on each bundle (sentence)

Each block must have exactly one entry point.

We'll now examine an example of a new block in file libs/blocks/Tutorial/Print_node_info.pm.

This block illustrates some of the most common methods for accessing objects:

• my @bundles = $document→get_bundles - an array of bundles contained in the document
• my $root_node = $bundle→get_tree($layer_name); - the root node of the tree of the given type in the given bundle
• my @children = $node→get_children; - array of the node's children
• my @descendants = $node→get_descendants; - array of the node's children and their children and children of their children …
• my $parent = $node→get_parent; - parent node of the given node, or undef for root
• my $root_node = $node→get_root; - the root node of the tree into which the node belongs

Attributes of documents, bundles or nodes can be accessed by attribute getters and setters, for example:

• $node→get_attr($attr_name);
• $node→set_attr($attr_name, $attr_value);

Our tutorial block Print_node_info.pm is ready to use. You only need to add this block to our scenario:

print_info:
        brunblocks -S -o Tutorial::Print_node_info -- sample.tmt

We can observe our new block behaviour:

make print_info

Try to change the block so that it prints out the information only for verbs. (You need to look at attribute tag at the m level). The tagset used is Penn Treebank Tagset.

It is assumed that finite clauses can be translated independently, which would reduce computational complexity, make parallel translation possible. We could even use hybrid translation - each finite clause could be translated by the most self-confident translation system.

A block which, given an analytical tree (SEnglishA), fills each a-node with boolean attribute is_head which is set to 1 if the a-node corresponds to a finite verb.

There is a block template with hints in libs/blocks/Tutorial/Mark_finite_clauses.pm. You should edit the block so that the ouput of this block is the same a-tree, in addition with attribute <is_head> attached to each a-node. There is also a printing block libs/blocks/Print_finite_clauses.pm which will print out the a-nodes grouped by clauses:

finite_clauses:
        brunblocks -S -o \
                Tutorial::Mark_heads \
                Tutorial::Print_finite_clauses -- sample.tmt

You are going to need these methods:

• $bundle→get_tree($tree_name)
• $node→get_attr($attr_name)
• $node→set_attr('attr_name',$attr_value)
• $node→get_eff_children()
• $node→get_children()

The output of our block is not ideal - we don't solve coordination and subordinate conjunctions. But for the time being, we can move on to some other tasks.

TODO

This time TODO vysvetleni a obrazek

You can use block template in devel/libs/blocks/BlockTemplate.pm. To see the results, you can again use TrEd (tmttred sample.tmt)

• TectoMT Developer's Guide - obsolete
• Questions? Ask kravalova at ufal.mff.cuni.cz
• Solutions to tasks in this tutorial are in libs/blocks/Tutorial/*solution.pm.
• TrEd - tree editor