Terms 0.1.0a2dev1

A smart knowledge store

The Terms knowledge store

Terms is a smart knowledge store. It is used to store knowledge, that can later be queried. It provides a declarative language with which to express the knowledge. It is smart because this language can be used to express rules, and these rules combine existing knowledge to produce new knowledge.

The Terms language

Here I will describe the Terms language.

To try the given examples, if you have installed Terms, you have to type “terms” in a terminal, and you will get a REPL where you can enter Terms constructs. Follow the instuctions in the INSTALL.rst.

More examples can be found in the github repository.

Words

The main building block of Terms constructs are words.

To start with, there are a few predefined words: word, verb, noun, number, thing, and exists.

New words are defined relating them to existing words.

There are 2 relations that can be established among pairs of words.

These relations are formally similar to the set relations “is an element of” and “is a subset of”.

In English, we express the first relation as “is of type”, and in Terms it is expressed as:

word1 is a word2.

So we would say that word1 is of type word2. The second relation is expressed in English as “is subtype of”, and in Terms:

a word1 is a word2.

So, we would say that word1 is a subtype of word2. Among the predifined words, these relations are given:

word is a word.
verb is a word.
a verb is a word.
noun is a word.
a noun is a word.
thing is a noun.
a thing is a word.
exists is a verb.
number is a word.

To define a new word, you put it in relation to an existing word. For example:

a person is a thing.
a man is a person.
a woman is a person.
john is a man.
sue is a woman.

These relations have consecuences, given by 2 implicit rules:

A is a B; a B is a C -> A is a C.
a A is a B; a B is a C -> a A is a C.

Therefore, from all the above, we have, for example, that:

thing is a word.
person is a word.
person is a noun.
john is a word.
a man is a thing.
john is a thing.
sue is a person.

With these words, we can build facts. A fact consists of a verb and any number of (labelled) objects.

Verbs are special words in that they have modifiers (or objects) when used to build facts. These modifiers are words, and are labeled. To define a new verb, you provide the types of words that can be objects for the verb in a fact, associated with their label. For example:

loves is exists, subj a person, who a person.

That can be read as: loves is a word of type verb, subtype of exists, and when used in facts it can take a subject of type person and an object labelled who of type person.

Facts

Facts are built with a verb and a number of objects. They are given in parenthesis. For example, we might have a fact such as:

(loves john, who sue).

The subj object is special: all verbs have it, and in sentences it is not labelled with subj, it just takes the place of the subject right after the verb.

Verbs inherit the object types of their ancestors. The primitive exists verb only takes one object, subj, of type word, inherited by all the rest of the verbs. So, if we define a verb:

adores is loves.

It will have a who object of type person. If adores had provided a new object, it would have been added to the inherited ones. A new verb can override an inherited object type to provide a subtype of the original object type (like we have done above with subj.)

Facts are not words, but they are also individuals of the language (terms), “first class citizens”, and can be used wherever a word can be used. Facts are of type exists, and also of type <verb>, were <verb> is the verb used to build the fact.

The objects in a fact can be of any type (a word, a verb, a noun, a thing, a number). In addition, they can also be facts (type exists). So, if we define a verb like:

wants is exists, subj a person, what a exists.

We can then build facts like:

(wants john, what (loves sue, who john)).

And indeed:

(wants john, what (wants sue, what (loves sue, who john))).

Rules

We can build rules, that function producing new facts out of existing (or newly added) ones. A rule has 2 sets of facts, the conditions and the consecuences. The facts in each set of facts are separated by semicolons, and the symbol -> separates the conditions from the consecuences. A simple rule might be:

(loves john, who sue)
->
(loves sue, who john).

The facts in the knowledge base are matched with the conditions of rules, and when all the conditions of a rule are matched by coherent facts, the consecuences are added to the knowledge base. The required coherence among matching facts concerns the variables in the conditions.

We can use variables in rules. They are logical variables, used only to match words or facts, and with a scope limited to the rule were they are used. We build variables by capitalizing the name of the type of terms that it can match, and appending any number of digits. So, for example, a variable Person1 would match any person, such as sue or john. With variables, we may build a rule like:

(loves Person1, who Person2)
->
(loves Person2, who Person1).

If we have this rule, and also that (loves john, who sue), the system will conclude that (loves sue, who john).

Variables can match whole facts. For example, with the verbs we have defined, we could build a rule such as:

(wants john, what (Exists1))
->
(Exists1).

With this, and (wants john, what (loves sue, who john))., the system would conclude that (loves sue, who john).

Variables that match verbs or nouns have a special form, in that they are prefixed by the name of a verb (or a noun), so that they match verbs that are subtypes of the given verb. For example, with the terms we have from above, we might make a rule like:

(LovesVerb1 john, who Person1)
->
(loves Person1, who john).

In this case, LovesVerb1 would match both loves and adores, so both (loves john, who sue) and (adores john, who sue) would produce the conclusion that (loves sue, who john).

For a more elaborate example we can define a new verb:

allowed is exists, subj a person, to a verb.

and a rule:

(wants Person1, what (LovesVerb1 Person1, who Person2));
(allowed Person1, to LovesVerb1)
->
(LovesVerb1 Person1, who Person2).

Then, (allowed john, to adores) would allow him to adore but not to love.

Likewise for noun variables. In this case an example might be PersonNoun1. This variable would match person, and also man and woman.

Finally, number variables are composed just with a capital letter and an integer, like N1, P3, or F122.

Numbers

Numbers, together with words and facts, are terms, and can be used wherever words or facts can. Numbers are of type number. We don’t define numbers, we just use them. Any sequence of characters that can be cast as a number type in Python are numbers in Terms: 1, -1e12, 2-3j, 10.009 are numbers.

Pythonic conditions

In rules, we can add a section where we test conditions with Python, or where we produce new variables out of existing ones. This is primarily provided to test arithmetic conditions and to perform arithetic operations. This section is placed after the conditions, between the symbols <- and ->. The results of the tests are placed in a condition python variable, and if it evaluates to False, the rule is not fired.

To give an example, let’s imagine some new terms:

aged is exists, age a number.
a bar is a thing.
club-momentos is a bar.
enters is exists, where a bar.

Now, we can build a rule such as:

(aged Person1, age N1);
(wants Person1, what (enters Person1, where Bar1))
<-
condition = N1 >= 18
->
(enters Person1, where Bar1).

If we have that:

(aged sue, age 17).
(aged john, age 19).
(wants sue, what (enters sue, where club-momentos)).
(wants john, what (enters john, where club-momentos)).

The system will (only) conclude that (enters john, where club-momentos).

Negation

We can use 2 kinds of negation in Terms, classical negation and negation by failure.

Classical negation

Any fact can be negated by prepending ! to its verb:

(!aged sue, age 17).

A negated fact is the same as a non-negated one. Only a negated fact can match a negated fact, and they can be asserted or used in rules. The only special thing about classical negation is that the system will not allow a fact and its negation if the same knowledge base: it will warn of a contradiction and will reject the offending fact.

Negation by failure

In pythonic conditions, we can use a function count with a single string argument, a Terms fact (possibly with variables), that will return the number of facts in the db matching the given one. We can use this to test for the absence of any given fact in the knowledge base, and thus have negation by failure.

Time

In the monotonic classical logic we have depicted so far, it is very simple to represent physical time: you only need to add a time object of type number to any temporal verb. However, to represent the present time, i.e., a changing distinguished instant of time, this logic is not enough. We need to use some non-monotonic tricks for that, that are implemented in Terms as a kind of temporal logic. This temporal logic can be activated in the settings file:

[mykb]
dbms = postgresql://terms:terms@localhost
dbname = mykb
time = normal

If it is activated, several things happen.

The first is that the system starts tracking the present time. It has an integer register whose value represents the current time. This register is updated each time we add new facts. There are 3 possible values for the mode setting for time: If the setting is none, nothing is done with time. If the setting is normal, the current time of the system is incremented by 1 when it is updated. If the setting is real, the current time of the system is updated with Python’s import time; int(time.time()).

The second thing that happens is that, rather than defining verbs extending exists, we use 2 new verbs, now and onwards, both subtypes of exists. These new verbs have special number objects: now has an at_ object, and onwards a since_ and a till_ objects.

The third is that the system starts keeping 2 different factsets, one for the present and one for the past. All reasoning occurs in the present factset. When we add a fact made with these verbs, the system automatically adds to now an at_ object and to onwards a since_ object, both with the value of its “present” register. The till_ object of onwards facts is left undefined. We never explicitly set those objects. When added, now facts go through the rule network, producing consecuences, and then are added to the past factset; onwards facts go through the rules network and then are added to the present factset. Queries for now facts go to the past factset, and those for onwards facts are done against the present. We might say that the facts in the present factset are in present continuous tense.

The fourth thing that happens when we activate the temporal logic is that we can use a new predicate in the consecuances of our rules: finish. We use it with an onwards fact: finish (<fact>). And when a rule with such a consecuence is activated, it grabs the provided fact from the present factset, adds to it a till_ object with the present time as value, removes it from the present factset, and adds it to the past factset. The system keeps track of the ancestry of facts obtained by reasoning, and when a fact is finished, its descent (if otherwise unsupported) is also finished.

Miscelaneous technical notes.

• I have shown several different kinds of variables, for things, for verbs, for numbers, for facts. But the logic behind Terms is first order, there is only one kind of individuals, and the proliferation of kinds of variables is just syntactic sugar. Person1 would be equivalent to something like “for all x, x is a person and x…”. LovesVerb1 would be equivalent to something like “for all x, a x is a loves and x…”.

• The design of the system is such that both adding new facts (with their consecuences) and querying for facts should be independent of the size of the knowledge base. The only place where we depend on the size of the data is in arithmetic conditions, since at present number objects are not indexed as such.

• The Python section of the rules is exec``ed with a dict with the ``condition variable in locals and an empty dict as globals. We might add whatever we like as globals; for example, numpy.

Installation and usage

Dependencies

Python 3 (tested on 3.2, 3.3).

Python libraries (these should be pulled by easy_install):

• ply

• sqlalchemy

Some RDBM system compatible with sqlalchemy (tested with postgreqsl and sqlite).

To run the tests, you need the nose framework.

Installation

It is advisable to install in a virtualenv.

If you have setuptools installed in your python, you can simply use easy_install, from a command line:

# easy_install Terms

Alternatively you can download the tarball, uncompress it, cd into the extracted directory, and run python3 setup.py install.

Interfacing

Once installed, you should have a terms script, that provides a REPL.

If you just type terms in the command line, you will get a command line interpreter, bound to an in-memory sqlite database.

If you want to make your Terms knowledge store persistent, You have to write a small configuration file ~/.terms.cfg:

[mykb]
dbms = sqlite:///home/eperez/mykbs
dmname = mykb
time = none

Then you must initialize the knowledge store:

$ initterms mykb

And now you can start the REPL:

$ terms mykb
>>>

In the configuration file you can put as many sections as you like, one for each knowledge store.

Installing Terms installs everything needed for the postgresql backend of sqlalchemy (except postgresql itself), so out of the box you can provide postgresql URLs in the config file.

The specified database must exist if you use postgresql, and the terms user (specified in the URL) must be able to create and drop tables and indexes.

So, for example, once you are set, open the REPL:

eperez@calandria$ initterms testing
eperez@calandria$ terms testing
>>> a person is a thing.
>>> loves is exists, subj a person, who a person.
>>> john is a person.
>>> sue is a person.
>>> (loves john, who sue).
>>> (loves john, who sue)?
true
>>> (loves sue, who john)?
false
>>> quit
eperez@calandria$ terms testing
>>> (loves john, who sue)?
true

Support

There is a mailing list at google groups. You can also open an issue in the tracker. Or mail me <enriquepablo at google’s mail domain>.