Automated Java Code Generation (ICDIM 2006)

Rule based Production Systems for Automatic Code Generation in Java

Imran Sarwar Bajwa1, M. Imran Siddique2 and M. Abbas Choudhary3
Department of Computer Sciecne and Information Technology
The Islamia University of Bahawalpur, Pakistan
bajwa@buitms.edu.pk

2 Facuty of Computer an EmergingSciences
3 Balochistan University of Information Technology and Management Sciences
Quetta, Pakistan.
imran@buitms.edu.pk
abbas@buitms.edu.pk

Abstract 1. Introduction
Unified modeling language is being used as a premier In the current age, the tools and techniques of software en-
tool for modeling the user requirements. These CASE gineering has been changed to an adequate extent. Now
tools provide an easy way to get efficient solutions. This every step of software engineering follows the rules of
paper presents a natural language processing based object oriented design patterns. Same the case is with Soft-
automated system for generating code in multi-languages ware process which uses Unified Modeling language for
after modeling the user requirements based on UML. UML modeling the user requirements. In recent times, there is no
diagrams are first generated by analyzing the given busi- software which provides services to draw UML diagrams
ness scenario provided by the user. A new model is presen- more efficiently except Rational Rose, Smart Draw etc and
ted for analyzing the natural languages and extracting the there is no doubt that these are reasonably good software
relative and required information from the given re- but has many drawbacks.
quirement notes by the user. User writes the requirements
in simple English in a few paragraphs and the designed Conventionally, the system analyst has to do a lot of work
system has conspicuous ability to analyze the given script. for deducing the business logic and understanding the user
After compound analysis and extraction of associated requirements before drawing the UML diagrams by using
information, the designed system draws various UML orthodox CASE tools. Hence, there is wastage of so much
diagrams as activity diagrams, sequence diagrams, class time due to the dull nature of the available CASE tools for
diagrams and Uses cases diagrams. The designed system the required scenario. In today’s world everybody needs a
has robust ability to create code automatically without ex- quick and reliable service. So it was needed that there
ternal environment. The designed system provides a quick should be some sort of intelligent software for generating
and reliable way to generate UML diagrams and generate UML based documentation to save time and budget of both
respective code to save the time and budget of both the the user and system analyst.
user and system analyst. In order to resolve all such issues, we need software, which
facilitates both users and software engineers. As far as this
software is concerns the time, it takes to explore all the fa-
cilities and services, should be quite less than a minute and
this information is quite useful for the users.

1-4244-0682-X/06/$20.00 ©2006 IEEE 300

Authorized licensed use limited to: The Islamia University of Bahawalpur. Downloaded on July 27, 2009 at 03:53 from IEEE Xplore. Restrictions apply.

2. Problem Description
The problem specifically addressed in this research is flavours as English language has more than half dozen
primarily related to the software analysis and design phase renowned flavours all over the world. These flavours have
of the software development process. Few years ago data different accents, set of vocabularies and phonological as-
flow diagram’s were being used to symbolize the flow of pects. These ominous and menacing discrepancies and in-
data and represent the user’s requirements. But in current consistencies in natural languages make it a difficult task
age, unified modeling language is used to model and map to process them as compared to the formal languages [13].
the user requirements, which is more comprehensive e and In the process of analyzing and understanding the natural
authentic way to of representation and it is beneficial for languages, various problems are usually faced by the
the later stages of software development. researchers. The problems connected to the greater
After modelling and mapping the user requirements, next complexity of the natural language are verb’s conjugation,
phase is to create the programming code in certain inflexion, lexical amplitude, problem of ambiguity, etc.
computer language. Hand written code by a programmer is From this set of problems the problem which ever causes
a conventional and orthodox solution of the problem which more difficulties is problem of ambiguity. Ambiguity could
is time consuming. Modern software engineering requires be easily solved at the syntax and semantic level by using a
quick and automated solutions which may have ability to sound and robust rule-based system.
create more than the half code, so that the programmer may
create the application after making appropriate adjustments 5. Object-Oriented Analysis and Design
and alterations in the automated generated code with less Analysis and design of an information system relates to un-
effort in less time as compared to the traditional derstand and intend the framework to accomplish the actual
approaches. job. Typically, design is relates to manage and control the
complexity parameter in a domain. A robust design
3. Problem’s Solution method also helps to split big tasks into controllable break-
The conducted research provides a robust solution to the ups (Condamines, 2001). In software engineering, design
addressed problem. Multi-lingual Code Generator (MCG) methods provide various notation usually graphical ones.
provides the solution of the problem. The functionality of These notations allow to store and communicate the perpetu-
the conducted research was domain specific but it can be al design decisions. Object-oriented design has overruled the
enhanced easily in the future according to the require- typical analysis and design techniques as structured design
ments. Current designed system incorporate the capability and data-driven design (Androutsopoulos, 1995). As com-
of mapping user requirements after reading the given pared to old style design paradigms, object-oriented design
models the every active entity of the problem domain using
requirements in plain text and drawing the set of UML dia-
concept of objects. Objects have:
grams as Class Diagram, Activity Diagram, Sequence
Diagram, Use case diagram and Component Diagram. State (shape and condition)
Behaviour (What they perform)
After drawing UML diagrams, designed system has pro-
found ability to create code in various languages as Java, Object-oriented languages use variable to manifest the state
Visual Basic, C # and C++. An Integrated Development of an object and methods or procedures to implement the
Environment has also been provided for User Interaction behaviour of an object. For example, a ball could be an ob-
and efficient Input and output. ject. There are different parameters of shape as colour, size,
diameter, shape, type, etc. This object can also have beha-
4. Natural Language Processing viour as throw, roll, catch, hit, etc. The major task in analysis
and design phase is to identify the valid objects and specify
The understanding and multi-aspect processing of the nat- there states and behaviours. In conventional methods, system
ural languages that are also termed as "speech languages", analyst performs this tough job and then maps this
is actually one of the arguments of greater interest in the information into UML using some graphical tool as Visio or
field artificial intelligence field [6]. The natural languages Rational Rose.
are irregular and asymmetrical. Traditionally, natural
languages are based on un-formal grammars. There are the In the context of this research, objects are automatically
geographical, psychological and sociological factors which identified from a problem domain. User provides the input
influence the behaviours of natural languages [17]. There text in English language related to the business domain.
are undefined set of words and they also change and vary After the lexical analysis of the text, syntax analysis is
area to area and time to time. Due to these variations and performed on word level to recognize the word category [4].
inconsistencies, the natural languages have differnet First of all the available lexicons are categorised i nto

301


nouns, pronouns, prepositions, adverbs, articles, conjunc-
Co-actor: The word with may introduce a join phrase
tions, etc. The syntactic analysis of the programs would
that serves a partner in the principal actor. They two
have to be in a position to isolate subject, verbs, objects,
carry out the action together "Zahid played tennis with
adverbs, adjectives and various other complements. It is
Ali."
little complex and multipart procedure.
"Zia is playing with the red ball." Recipient: The recipient is the person for whom an
action has bee performed: "Ali bought the balls for
For this example, following is the output. Zahid." In this sentence Suzie is beneficiary.
Lexicons Phase-I Phase –II Thematic object: The thematic object is the object the
sentence is really all about. Often the thematic object is
Zia Noun Object
the same as the syntactic direct object, as "Zahid hit the
is Helping-Verb -------
ball." Here the ball is thematic object.
playing Verb Method
with Preposition ------- Transportation: The transportation is something in
the Article ------- which or on which travels: “Aslam always goes by
red Noun Attribute train."
ball Noun Object
Path: Motion from source to destination takes place
This is the final output of lexical assessment phase and over at path. Several prepositions can serve to introduce
all nouns are marked as objects and verbs are marked as trajectory noun phrases: "Ali and Zahid went to
methods and all adjective are marked as states of that par- Islamabad through Lahore."
ticular object. In the above example, there are two objects
‘Zia’ and ball. ‘Playing’ is method of ‘Zia’ and ‘red’ is the Position: The position is where an action occurs. As in
concerned attribute of the object ‘ball’. the path role, "several prepositions are possible, which
conveys means in addition of serving as a signal that a
position noun phrase is "Aslam and Ahmed studied in
6. Used Methodology the library, at a desk, by the wall, near the door."
Conventional natural language processing based systems
Period: Period specifies when an action occurs.
user rule based systems. Agents are another way to address
Prepositions such at, before, and after introduce noun
this problem[8]. In the research, a rule-based algorithm has
phrases serving as time role fill "Ahmed and Ali left
been used which has robust ability to read, understand and
before noon."
extract the desired information. First of all basic elements
of the language grammar are extracted as verbs, nouns, ad- Interval: Interval specifies how long an action takes.
jectives, etc then on the basis of this extracted information Preposition such as for indicate duration. "Aslam and
further processing is performed. In linguistic terms, verbs Ahmed jogged for an hour.”
often specify actions, and noun phrases the objects that
participate in the action [16]. Each noun phrase's then role 7. Architecture of Designed System
specifies how the object participates in the action. As in the
The designed UMLG system has ability to draw UML
following example:
diagrams after reading the text scenario provided by the
"Robbie hit a ball with a racket." user. This system draws diagrams in five modules: Text
input acquisition, text understanding, knowledge extrac-
A procedure that understands such a sentence must dis-
tion, generation of UML diagrams and finally multi-lingual
cover that Role is the agent because he performs the action
code generation as shown in following fig.
of hitting, that the ball as the thematic object because it is
the object hit, and that the racket is an instrument because Text
it is the tool with which hitting is done. Thus, sentence
analysis requires, in part, the answers to these actions: The Text Input Acquisition
number of thematic roles embraced by various theories
varies probably. Some people use about half-dozen
thematic roles [9]. Others use more times as many. The ex-
act number does not matter much, as long as they will great Text Understanding
enough to expose natural constraints on how verbs and
thematic-instances form sentences.
Actor: The actor causes the action to occur as in "Zahid
Knowledge Extraction
hits the ball," Zahid is actor who performs the task.

302


UML Diagrams
7.4. UML diagram generation
This module finally uses UML symbols to constitute the
various UML diagrams by combining available symbols
Multi-lingual Code Generation
according to the information extracted of the previous
module. As separate scenario will be provided for various
Code diagrams as classes, sequence, activity and use cases dia-
grams, so the separate functions are implemented for re-
Figure 1: Architecture of the designed MCG system spective diagram.

7.1. Text input acquisition
This module helps to acquire input text scenario. User
provides the business scenario in from of paragraphs of the
text. This module reads the input text in the form
characters and generates the words by concatenating the
input characters. This module is the implementation of the
lexical phase. Lexicons and tokens are generated in this
module.
7.2. Text Understanding
This module reads the input from module 1 in the from of
words. These words are categorized into various classes as
verbs, helping verbs, nouns, pronouns, adjectives, preposi-
tions, conjunctions, etc.
7.3. Knowledge extraction
This module, extracts different objects and classes and their
respective attributes on the basses of the input provided by
the preceding module. Nouns are symbolized as classes and Figure 3: Generating class diagrams
objects and their associated attributes are termed as
attributes. 7.5. Multi-lingual Code generation
This is the last module, which ultimately generate code in
the different popular languages as Java, C#.Net and
VB.Net to help the programmer. The generated code is
structured according to the knowledge extracted in the pre-
vious modules and the UML diagrams generated.

Figure 2. Extracting classes, functions and attributes Figure 4: Generating Code in Java Language

303


8. Accuracy Evaluation that he has followed the prerequisites of the software to
prepare the input scenario. The given scenario should be
To test the accuracy of the diagrams generated by the complete and written in simple and correct English. Under
designed system four parameters had been decided. Each the scope of our project, software will perform a complete
generated diagram from each category was checked. Max- analysis of the scenario to find the classes, their attributes
imum score was declared 25. According to the wrong nom- and operations. It will also draw the diagrams as class dia-
inations and extractions, the points were detected. A matrix grams, activity diagrams, use-case diagrams and sequence
of results of generated diagrams is shown below. diagrams.

Table 1 - Testing results of different UML Diagrams An elegant graphical user interface has also been provided
to the user for entering the Input scenario in a proper way
and generating UML diagrams.
Dig. Types Objects Attributes Sequence labeling Total

Class 22 24 20 19 85% 10.Future Work
Activity 23 21 16 20 80% The designed system for generating UML diagrams and
their respective code in multiple languages was started
Sequence 21 22 13 18 74%
with the aims that there should be a software which can
Use case 21 24 21 22 88% read the scenario given in English language and can draw
the all types of the UML diagrams such as Class diagram,
activity diagram, sequence diagram, use case diagram,
A matrix representing UML diagrams accuracy test (%)
component diagram, deployment diagram. But last two of
for class, activity, sequence and use case diagrams has
them component diagram, deployment diagram are still un-
been constructed. Overall diagrams accuracy for all types
touched.
of UML diagrams is determined by adding total accuracy
of all categories and calculating average of it. Following There is also some margin of improvements in the
graph is showing the accuracy ratio of various diagram algorithms for generating first four types Class diagram,
types in terms of objects, attributes, sequence and labeling activity diagram, sequence diagram, use case diagram.
parameters. Current accuracy of generating diagrams is about 80% to
85%. It can be enhanced up to 95% by improving the
algorithms and inducing the ability of learning in the sys-
tem.

11.References
[1] Allen,J. (1994) Natural Language Understanding.
Benjamin- Cummings Publishing Company, New York.
[2] Biber, D., Conrad, S., & Reppen, R. (1998). Corpus
Linguistics: Investigating Language Structure and Use.
Cambridge Univ. Press, Cambridge, U.K.
[3] Blaschke,C., Andrade,M.A., Ouzounis,C. and
Valencia,A. (1999) Automatic extraction of biological
Figure 5: Accuracy Ratio of various diagram types information from scientific text: protein–protein
interactions. Ismb, 60–67.
9. Conclusion [4] C. A. Thompson, R. J. Mooney and L. R. Tang,
Learning to parse natural language database queries into
This research is all about the dynamic generation of the logical form, in: Workshop on Automata Induction,
UML diagrams and their respective code by reading and Grammatical Inference and Language Acquisition (1997).
analyzing the given scenario in English language provided by
the user. The designed system can find out the classes and [5] Chomsky, N. (1959). On certain formal properties of
objects and their attributes and operations using an artificial grammars. Information and Control, 2(2), 137–167.
intelligence technique such as natural language processing. [6] Chomsky, N. (1965). Aspects of the Theory of Syntax. MIT
Then the UML diagrams such as Activity dig., Sequence Press, Cambridge, Mass. Chow, C., & Liu, C. (1968).
dig., Component dig., Use Case dig., etc would be drawn. Approximating discrete probability distributions with
The accuracy of the software is expected up to about 80% dependence trees. IEEE Transactions on Information Theory,
with the involvement of the software engineer provided IT-14(3), 462–467.

304


[7] Fagan, J. L. (1989). The effectiveness of a non-syntactic
approach to automatic phrase indexing for document retrieval.
Journal of the American Society for Information Science, 40(2),
115–132.
[8] J. M. Zelle and R. J. Mooney, Learning semantic
grammars with constructive inductive logic programming,
in: Proceedings of the 11th National Conference on
Arti_cial Intelligence (AAAI Press/MIT Press,
Washington, D.C., 1993), pp. 817ñ822.
[9] Kowalski, G. (1998). Information Retrieval Systems: Theory
and Implementation. Kluwer, Boston.
[10] Krovetz, R., & Croft, W. B. (1992). Lexical ambiguity and
information retrieval. ACM Transactions on Information
Systems, 10, 115–141.
[11] Losee, R. M. (1988). Parameter estimation for probabilistic
document retrieval models. Journal of the American Society for
Information Science, 39(1), 8–16.
[12] Losee, R. M. (1996a). Learning syntactic rules and tags
with genetic algorithms for information retrieval and filtering:
An empirical basis for grammatical rules. Information
Processing and Management, 32(2), 185–197.
[13] Manning, C. D., & Schutze, H. (1999). Foundations of
Statistical Natural Language Processing. MIT Press,
Cambridge, Mass.
[14] Maron, M. E.,& Kuhns, J. L. (1960). On relevance,
probabilistic indexing, and information retrieval. Journal of the
ACM, 7, 216–244.
[15] Partee, B. H., Meulen, A. t., &Wall, R. E. (1990).
Mathematical Methods in Linguistics. Kluwer, Dordrecht, The
Netherlands.
[16] Salton, G., & McGill, M. (1983). Introduction to Modern
Information Retrieval. McGraw-Hill, New York.
[17] S. Weiss, C. Apte, F. Damerau, D. Johnson, F. Oles,
T. Goetz and T. Hampp, Maximizing text-mining
performance, IEEE Intelligent Systems 14 (1999) 63ñ69.
[18] Strzalowski, T. (1995). Natural language information
retrieval. Information Processing and Management, 31(3), 397–
417.
[19] Van Rijsbergen, C. (1977). A theoretical basis for use of
co-occurrence data in information retrieval. Journal of
Documentation, 33(2), 106–119.

305


Automated Java Code Generation (ICDIM 2006)

More Related Content

What's hot (20)

Similar to Automated Java Code Generation (ICDIM 2006) (20)

More from IT Industry (14)

Automated Java Code Generation (ICDIM 2006)