A PML-TQ query consists of a selective part that selects nodes from the treebank and an optional sequence of output filters that are used to extract data from the matching nodes, post-process the results, compute statistics, generate tabular output, etc.

The selective part of a PML-TQ query postulates requirements on one or more nodes from the treebank and their mutual relationships (e.g. on the topological configuration in the tree structure). It is formed by one or more node selectors, which form the outermost scope of the query. Inner scopes of the query are given by nested subqueries as described later.

A node selector represents a node in the treebank of a certain type (in the PML data model, the nodes in the treebank annotation can be typed; the query can also refer to several annotation layers with different types of nodes) and postulates constraints on its properties including relationships to nodes represented by other selectors.

Selectors may nest other selectors; a nested selector belongs to the same scope as the containing selector The nested selector may explicitly specify the relation of its matching node to the node matched by the containing selector; the default relation is child. The nesting of selectors can thus naturally follow the topology of the matching tree.

Selectors can also be named and referred to from other node selectors; however, in many cases, the need for explicitly naming them can be eliminated by nesting.

A match of a query is a mapping which assigns to each outermost-scoped selector a node from a treebank (called a matching node) of the type specified by the selector, in such a way that all the matching nodes are mutually distinct and simultaneously satisfy the constraints postulated by their corresponding selectors (including constraints on their mutual relationships). The match can be represented as a tuple of the matching nodes ordered accordingly to some canonical ordering of the selectors from the outermost scope of the query. There can be zero, one, or more distinct matches of the query in the treebank (two matches are distinct if, as ordered tuples, they differ in at least one node).

Non-identity rule: Two distinct selectors in the same scope of the query always represent two distinct nodes in each match of the query or sub-query (unless explicitly specified othewise in the query).

Selectors can postulate the following types of constraints:

In the following descriptions, we refer to the selector postulating a constraint as as the current selector.

Predicate constraints assert equality, inequality, or regular expression match between values computed from terms. An atomic term is a constant (integer, float, or character string), or an attribute of a node matched by the current selector or some other selector in the current or outer scope of the query. A term is either an atomic term or a term obtained from other terms using arithmetical (+, *, -, div, mod) or string (concatenation & ) operators, or functions.

A reference is a constraint on the relationship of a node matched by some named selector to the node matched by the current selector. The referred selector must either belong to the same scope as the current selector or to its outer scope.

A subquery is formed by a selector (called the leading selector of the subquery) nested in the current selector and augmented by restrictions on the number of occurrences, computed as the number of distinct nodes matched by the leading selector of the subquery relatively to a fixed match of the selectors in the current and outer scope (including the current selector). For example, to postulate a constraint that each node matched by the current selector must have at least two child nodes, we create a subquery in form of a nested selector in the child relation to the current selector and restrict the number of occurrences to two and more.

The leading selector can nest other selectors. Each subquery starts a new scope whose outer scope is the scope of the containing selector together with the containing selector's outer scope (if any). Unlike selectors from the outermost scope, selectors declared within a subquery do not represent any particular node in the resulting match. They can refer to selectors from the same scope, and also to selectors from the outer scope, but not vice versa (selectors from the outer scope cannot refer to the selectors in the subquery).

A subquery constraint is verified as follows: for each match of the selectors in the current and outer scope, all matches of the subquery are located (these may coincide with nodes matched by the selectors in the outer scope). The number of distinct nodes matched by the leading selector of the subquery are counted and this number is compared with the restrictions on number of occurrences. The constraint is satisfied if and only if these restrictions are met.

A constraint can also be a boolean combination of other constraints; a nested node selector occurring in a boolean combination with other nested node selectors or constraints is considered to be a subquery with at least one occurrence.

A PML-TQ query can be visualized as a graph consisting of one or more trees whose nodes are the selectors connected by edges according to the nesting of selectors and subqueries. In this sense we may sometimes refer to selectors as query nodes and to the query as query graph or query tree (a technical root can be added above all the trees so that a forest becomes a single tree). The edges can be labeled or colored to represent different relationships between nodes. References to named selectors can be represented by an additional layer of links (edges) in the graph that may go across the basic tree structure of the query tree.

The PML-TQ provides a client interface in the form of a web application that can be accessed by any web browser capable of combining JavaScript, CSS, and SVG (Scalable Vector Graphics), such as Firefox, Google Chrome, Opera browser, and Safari.

Unlike the TrEd interface, this interface does not require any installation, but lacks some features such as graphical query builder and graphical representation of the query (the queries must be entered in the text form), and of course does not support querying local files.

To access the PML-TQ servers hosted at Lindat/Clarin servers, go to lindat.cz and in the top menu, click on  . A starting PML-TQ web page will be displayed (see Figure 1, “The PML-TQ web-service at Lindat/Clarin web pages”).

Figure 1. The PML-TQ web-service at Lindat/Clarin web pages

Below, you can see two lists – a list of recently used treebanks and a list of featured treebanks. Clicking on any of the listed treebanks will connect you directly to the server for the given treebank.

A list of all available treebanks can be accessed by clicking on  , as demonstrated in Figure 2, “The list of available treebanks at Lindat/Clarin web pages”.

Figure 2. The list of available treebanks at Lindat/Clarin web pages

Here, you can filter the treebanks according to their public availability (i.e. accessibility of the server without login), language and other tags. By clicking on a single treebank from the (filtered) list below, you will get connected to the search server for the respective treebank. For example, if you select the Prague Dependency Treebank 3.0 (PDT 3.0), you will get to the following web page (Figure 3, “The help page for the PDT 3.0 treebank”):

Figure 3. The help page for the PDT 3.0 treebank

It is a very short introduction to the query language that should help you start searching in the treebank if you do not wish to read through this lengthy tutorial. To proceed to the page where you can actually enter a search query, click on  .

The following web page will be displayed (Figure 4, “The start page for searching in the PDT 3.0 treebank”):

Figure 4. The start page for searching in the PDT 3.0 treebank

Now we may create our first simple query. We shall search for all nodes of the type t-node (tectogrammatical nodes in PDT 3.0) whose attribute functor equals to PRED (Predicate). In the web client, the query can be created in two ways:

To execute the query, press   below the query text area. The query gets processed by the server and the result is displayed. Figure 6, “A simple query searching for Predicates in PDT 3.0” shows the first matching tree.

Figure 9. A result tree for the query searching for Predicates in PDT 3.0

The corresponding sentence is displayed just above the tree. Try clicking on the individual words of the sentence and see the animation marking the corresponding nodes in the tree.

Buttons and can be used to navigate among the results, and buttons and to see context sentences/trees. To go directly to the N-th result, change the number of the current result () to a desired number (make sure that the focus is in the result number field) and – in the list of matching nodes – click on . The corresponding matching node in the given result is displayed (and highlighted in the same colour, in this case green).

We shall now make the query more complex by adding another node to it. We shall ask for a t-node with functor "PRED" (Predicate) that has a child with functor "PAT" (Patient).

To add a node to an existing one in the query, you need to specify a type of relation of the new node to the existing one. The list of available relations can be accessed through in the toolbar (see Figure 10, “A part of the list of available standard relations between nodes in PML-TQ”).

Figure 10. A part of the list of available standard relations between nodes in PML-TQ

The default value is child, so the following two queries are equivalent:

t-node [ functor="PRED", t-node [ functor="PAT" ] ]

try the query

or

t-node [ functor="PRED", child t-node [ functor="PAT" ] ]

try the query

These forms of the query use nesting of node selectors. The query can also be expressed without nesting, using names, either as

t-node $a := [ functor="PRED", child $b ];
t-node $b := [ functor="PAT" ];

try the query

or

t-node $a := [ functor="PRED" ];
t-node $b := [ functor="PAT", parent $a ];

try the query

naming the two nodes $a and $b and either indicating that $a has a child $b or that $b has a parent $a.

We now extend our query to cover t-nodes not only with functor PAT (Patient) but also ADDR (Addressee). This can be done in three different ways:

Using a disjunction:

t-node [ functor="PRED", t-node [ functor="PAT" or functor="ADDR" ] ]

try the query

Using a regular expression:

t-node [ functor="PRED", t-node [ functor ~ "^(PAT|ADDR)$" ] ]

try the query

Symbol ~ (tilde) denotes a binary relation between two values that is true if and only if the value on the left interpreted as string matches the value on the right interpreted as regular expression.

Using a set enumeration:

t-node [ functor="PRED", t-node [ functor in { "PAT", "ADDR" } ] ]

try the query

The relation in asserts that the value computed from the expression on the left equals to a value of some of the expressions listed in the set enumeration on the right.

The member selector is useful for querying some types of complex-valued node attributes, e.g. lists of complex structures. In PDT 3.0, there are three such attributes, representing discourse relations (attribute discourse), textual coreference (coref_text) and bridging anaphora (bridging).

For example, attribute bridging is a list of structures, each consisting of two members: a PML reference to the anaphor node (bridging/target_node.rf) and the type of the relation (bridging/type). So, each structure in the list represents one labeled semantic relation. To be able to combine constraints on the target node (anaphor) with the type of the relation that points to it, we must use a feature of PML-TQ called member selectors.

The following query searches for bridging relations of type PART_WHOLE.

t-node [
  member bridging [
    type = "PART_WHOLE",
    target_node.rf t-node [ ]
  ]
]

The intermediate member selector matches one element of the bridging list at a time and tests its type. If the type matches, the nested node selector for the target_node.rf PML-reference relation takes action.

Similarly, we can search for discourse relations of a given discourse type. The following example searches for discourse relations of discourse type opp (opposition). The query also specifies that the target node is in the same tree as the start node, thus searching only for intra-sentential discourse relations.

t-node $s:= [
  member discourse [
    discourse_type = "opp",
    target_node.rf t-node [ same-tree-as $s ]
  ]
]

Sometimes it is useful to test existence, non-existence or number of occurrences of a node related to our query. For example, to find all predicates without a subject on the analytical layer of PDT 3.0, we could use the following query

a-node [ afun="Pred", 0x echild a-node [ afun="Sb" ] ]

The query finds an a-node with afun="Pred" that has no effective children with afun="Sb". This is expressed using a selector preceded by a restriction on number of occurrences (0x - zero times), which is called a subquery.

Of course, we could constrain the number of occurrences to a non-zero value, too. For example, to find all predicates that govern one subject or one object, but not both, we could use the following query:

a-node [ afun="Pred", 1x echild a-node [ afun in {"Sb","Obj"} ] ]

The nodes matched by subqueries are not part of the result match (in our example, the match would consist of the predicate nodes only, the subjects or objects would not be included).

The number of occurrences of a subquery can be constrained not only to a single number but to any finite union of bounded or partially unbounded intervals of positive integers; e.g. 0|2..4|6+x restricts the number of occurrences to zero, two to four, or six or more, eliminating one and five. While the plus sign stands for or more, the minus sign means or less, as in 4-x (occurring four or less times).

Subqueries are also created using boolean operators, such as negation:

a-node [ afun="Pred", ! echild a-node [ afun="Sb" ] ]

In this example, the selector ! echild a-node [ afun="Sb" ] is automatically turned into a (still negated) subquery with one and more occurrences; the query becomes:

a-node [ afun="Pred", ! 1+x echild a-node [ afun="Sb" ] ]

try the query

A common use of subqueries is also constraining nodes on a descending path from one node to another. Let us for example formulate a query searching for a descending chain of tectogrammatical nodes with the functor RSTR (restrictive or descriptive abdominal modification). We want the chain to satisfy the following conditions:

The corresponding query looks like this:

t-node $N:= [
  # condition 1.
  gram/sempos ~ "^n",
  functor != "RSTR",

  # conditions 2. and 3.
  descendant{3,} t-node $R := [
    functor = "RSTR",
    # condition 4.
    0x  t-node [ functor = "RSTR" ]
  ],

  # condition 5.
  0x descendant t-node [
    !functor = "RSTR",
    descendant $R
  ],
];

Note how the condition 5. is expressed: we say that there is no descendant of $N dominating $R whose functor would not equal RSTR. Thus, we have rewritten the original condition of the form ∀x C(x,N,R) as Ź ∃x ŹC(x,N,R).

Sometimes you want to find a good small example tree demonstrating some linguistic phenomenon. You want it to fit to a presentation slide or an article page. You can do so by putting a limit on the tree size.

Using a subquery this can be done as follows:

t-node [
  10-x same-tree-as t-node [],
  functor="DPHR", # the rest of your query
]

This selects t-nodes with functor="DPHR" in trees with at most 10 other t-nodes. Using functions (see below), this can be written as

t-root [
  descendants() <= 10,
  descendant t-node [ functor="DPHR" ]
]

but note that in this case the t-root appears as a node in the result set. To avoid it, we can write

t-node [
  functor="DPHR",
  1x ancestor t-root [ descendants() <= 10 ]
]

For treebanks that do not have a special node type for the root node, we can write e.g.:

node [
  functor="DPHR",
  1x ancestor node [
    depth() = 0, # the root
    descendants() <= 10
  ]
]

PML-TQ provides a set of built-in functions that can be used in expressions constraining nodes and also in output filters. The functions can be split into the following categories:

In the web client, lists of available functions can be accessed via in the toolbar, see Figure 11, “A part of the list of available functions in PML-TQ”.

Figure 11. A part of the list of available functions in PML-TQ

For description of all individual functions, refer to the PML-TQ manual. Here, we only give a few examples demonstrating the use of some of the functions from the first category on a few common query constructions, usually also expressible by means of subqueries. Whether it is more efficient to use functions than subqueries may depend on implementation.

# a leaf node (using functions)
t-node [ sons()=0 ]

# a leaf node (using a subquery)
t-node [ 0x child t-node [ ] ]

# right-most child
t-node [ rbrothers()=0 ]

# left-most child
t-node [ lbrothers()=0 ]

# first leaf node in a subtree of $t (using functions)
t-node $t := [
  descendant t-node [
    sons()=0,
    depth_first_order()-depth()=depth_first_order($t)-depth($t)
  ]
]

# first leaf node in a subtree of $t (using a subquery)
t-node $t := [
  descendant t-node $d := [ sons()=0 ],
  0x descendant [ sons()=0, depth-first-precedes $d ],
]

# last leaf node in a subtree of $t
t-node $t := [
  descendant t-node [
    sons()=0,
    depth_first_order()-depth()=depth_first_order($t)+descendants($t)-1-depth($t)
  ]
]

# last leaf node in a subtree of $t (using a subquery)
t-node $t := [
  descendant t-node $d := [ sons()=0 ],
  0x descendant [ depth-first-follows $d ],
]

Output filters are used for extracting data from the nodes matched by the query and generating tabular output. Filters must follow the selective part of the query and start with >>. Filters can be chained: the first filter extracts data from the matching nodes and all subsequent filters operate on the output from the immediately preceding filter. Details can be found in the PML-TQ manual, sections PML-TQ Syntax Reference and Group Functions.

One of the simplest filters uses the group function count() to compute the total number of matches of the query in the treebank:

# counting occurrences
t-node [ functor="PRED" ]
>> count()

The group functions min(), max(), and avg(), can be used to compute maximum, minimum, and average values of data extracted from the matching nodes. For example: to compute a maximum number of child nodes of a t-node with the functor PRED, we can use the following:

t-node $n := [ functor="PRED" ]
>> max(sons($n))

The following query computes maximum, minimum and average size of a tectogrammatical tree:

t-root $n := [ ]
>> descendants($n)
>> max(), min(), avg()

The above query uses two filters: the first extracts the number of descendants from each node matched by the selector $n, the second computes maximum, minimum and average value from the values returned by the first filter.

The following query shows a common grouping construction using the 'for' clause. It extracts the attribute functor from the matched nodes and for each distinct value counts the number it occurred:

t-node $n := [ ]
>> for $n.functor
   give $1, count()

Note that $1 in the give clause refers to the first (and only) key used in the for clause, i.e. to $n.functor.

By appending a sort by clause to a filter, we may reorder the rows it produces by some of its columns. In the following query, the output of the filter is sorted using the second output column (the count()) in descendant order as the primary key and the first output column (the $1 in the give clause) in the default (ascending) order:

t-node $n := [ ]
>> for $n.functor
   give $1, count()
   sort by $2 desc, $1

The for clause can be used to create groups not only by attribute values, but also by some of the matching nodes. For example, in order to find out how many grammatical-coreference links can start in one tectogrammatical node, we may use the following query:

t-node $referring := [
  coref_gram.rf t-node $referred := [  ]
];
>>  for $referring give count()
>> max()

The selective part of the query matches every pair of tectogrammatical nodes that are linked by a grammatical-coreference link. The first filter groups the resulting pairs of nodes by the first of the nodes ($referring) and outputs the number of pairs in each group; this is the number of grammatical-coreference links starting in the node $referring. The second filter simply computes the maximum of the values returned by the first filter.

The for clause partitions all input rows into groups before any further processing and the subsequent give clause then produces one output row for each group, letting all group functions, such as count(), min(), max(), etc. operate on the particular group.

PML-TQ further supports a syntax that allows different partitions to be defined for different group function and also let the give clause operate on all input rows. This is done by following the function arguments by an over clause. Here we show an example where we use one of the ranking group functions (row_number()) to select just a few top ranking rows from each group. Please refer to the PML-TQ manual, section Group functions explained for more examples.

In the following query we extract the syntactic label (afun) and the part of speech (the first position of the morphological tag) from every node on the analytical (morphosyntactical) layer of PDT 3.0. Then we apply further filters to output in order to obtain the three most frequent parts of speech for each afun. If several parts of speech occur the same number of times for a given afun, we sample those three that come first alphabetically.

a-node $a:= [ ]
>> $a.afun, substr($a.m/tag,0,1)  # get afun and part of speech (POS)
>> for $1,$2 give $1, $2, count() # count occurrences of POS for each afun
>> $1, $2, row_number(over $1 sort by $3 desc, $2) # get the rank of each POS over the afun
   sort by $1, $3
>> filter $3 <= 3
>> $1, $2, $3