Search

Magistrska naloga predstavlja zasnovo in implementacijo orodja Revertium, ki uporabniku pomaga pri identifikaciji možnosti optimizacije relacijskega podatkovnega modela na podlagi analize dnevnika transakcij obstoječe relacijske podatkovne baze. Raziskali smo relacijski podatkovni model in določene slabosti relacijskih podatkovnih baz ter na podlagi pridobljenega znanja definirali možnosti optimizacije le-teh. Za namen doseganja prej omenjenega smo raziskali povratni inženiring relacijskih podatkovnih baz in določene optimizacijske tehnike, kot sta indeksiranje in denormalizacija. Na podlagi ugotovljenega smo združili vso znanje v celoto – orodje Revertium, ki je rezultat magistrskega dela. The thesis represents the design and implementation of the tool Revertium, which helps the user identify the possibilities of optimizing the relational data model based on the analysis of database transaction logs. We have researched the relational data model and its weaknesses to define the optimization options based on acquired knowledge. For the purpose of achieving the latter, we researched reverse engineering of relational databases and some optimization techniques like indexing and denormalization. Based on all acquired knowledge, we created a proposed result of the thesis – a tool named Revertium.

V magistrskem delu se spoznamo s spletnimi iskalniki, njihovim delovanjem in tehnikami za njihovo izdelavo. Poskušamo razviti lasten spletni iskalnik, ki temelji na tehnikah in tehnologijah, opisanih v teoretičnem delu. Spletni iskalnik uporabnikom omogoča iskanje po rezultatih, ki v ozadju neprestano posodabljajo svoj indeks s pomočjo spletnih pajkov. Uporabnost izdelane rešitve pokažemo tudi na dveh primerih. Ugotovili smo, da je v sklopu magistrskega dela preveč omejitev, da bi lahko izdelali rešitev, ki bi iskala po globalnih rezultatih in se približala iskalnikom, kot je Google. Lahko pa izdelamo uporaben lokalni iskalnik, ki išče po določeni spletni strani. Takšen se je med testiranji odrezal celo bolje od iskalnika Google. In this master's thesis, we get to know search engines, how they function and what techniques can be used to develop our search engine which bases on techniques and technologies described in the theoretical part. A search engine allows users to search results which are continually being updated in the background using web crawlers. We also show the usefulness of the created solution on two examples. We found out that within the scope of the thesis there are too many restrictions to be able to produce a solution that would use global search data and be as good as Google. What we can create however is a local search engine that is only limited to a particular website which in tests showed to be even better than Google.

V okviru diplomske naloge smo izdelali novo rešitev za evropsko digitalno knjižnico Europeana. Razvita rešitev preko spletnega programskega vmesnika omogoča evropskim kulturnim ustanovam veliko hitrejšo objavo svojih zapisov oziroma del, v primerjavi s trenutnim procesom, na spletni portal evropske digitalne knjižnice Europeana. Spletni programski vmesnik omogoča tudi zelo hitro obdelavo podatkov, ne glede na že prisotno količino le-teh. Prav tako smo na tej veliki zbirki podatkov implementirali tekstovno iskanje v skoraj realnem času ter tako omogočili čim hitrejši dostop do informacij preko spletnega portala Europeana. As part of this thesis we have developed a new solution for European digital library – Europeana. Newly developed solution allows European cultural institutions to post their publications on Europeana must faster, compared to the current process. We have developed a new web API which provides much faster data processing regardless of the amount of the current existing data. We also added support for full-text search in near real time and therefore enabling users to quickly find information about posted publications on Europeana portal.

Uporabniki morajo znati identificirati najrelevantnejše vire kakovostnih znanstvenih in tehničnih informacij. Razumeti morajo strukturo zbirk, iskalne tehnike, kontroliran besednjak, terminologijo in klasifikacijske sheme. V članku podajamo glavne značilnosti bibliografskih podatkovnih zbirk (Agricola,Agris, Cab Abstracts (CABI), Compendex, Inspec, Iconda), statističnih zbirk (Eurostat, Faostat) in zbirk tehničnih dokumentov - standardov in patentov (ISO, WIPO) glede na značilnosti lesarstva oz. predelave gozdnih proizvodov. Opisane so razlike v posameznih iskalnih poljih in iskalna načela (platforme, Boolovi logični operatorji, predpone, pripone, znaki za krajšanje ali maskiranje). Posebno pozornost namenjamo predmetnim oznakam: ožje ključne besede (Descriptors, Heading Words, Controlled Terms, Keywords) in širše predmetne kategorije (Subject Categories, Concept Codes, CABI Codes, Classification Codes). Podani so seštevki besed oz. indeksnih gesel ali pojmov z najvišjo pojavnostjo. S tem pregledom želimo spodbuditi rabo dodatnih podatkovnih zbirk ter rabo strukturiranih besednjakov, kar naj se izrazi v boljših iskalnih rezultatih. End-users should be able to identify the most relevant scientific and technical information resources. They must understand database structure, query techniques, controlled dictionaries, terminology and indexing schemes. In the present article, we review principal characteristics of the following information systems: bibliographic databases (Agricola, Agris, Cab Abstracts (CABI), Compendex, Inspec, Iconda, statistical collections Eurostat and Faostat, and technical standards and patents ISO and WIPO with regard to the field of wood sciences or processing of forest products. Differences in database fields and search characteristics (search platforms, Boolean logic, prefixes, suffixes, wildcards) are described. Special attention is placed on subject headings in each respective database: narrow keywords (Descriptors, Heading Words, Controlled Terms, Keywords) and broader categories (Subject Categories, Concept Codes, CABI Codes, Classification Codes). Word counts are shown for the terms with the highest frequency of occurrences. This review promotes the use of additional databases, encourages employment of structured vocabularies, which should result in improved search results.

Uporabniki morajo znati identificirati najrelevantnejše vire kakovostnih znanstvenih in tehničnih informacij. Razumeti morajo strukturo zbirk, iskalne tehnike, kontroliran besednjak, terminologijo in klasifikacijske sheme. V članku podajamo glavne značilnosti bibliografskih podatkovnih zbirk (Agricola,Agris, Cab Abstracts (CABI), Compendex, Inspec, Iconda), statističnih zbirk (Eurostat, Faostat) in zbirk tehničnih dokumentov - standardov in patentov (ISO, WIPO) glede na značilnosti lesarstva oz. predelave gozdnih proizvodov. Opisane so razlike v posameznih iskalnih poljih in iskalna načela (platforme, Boolovi logični operatorji, predpone, pripone, znaki za krajšanje ali maskiranje). Posebno pozornost namenjamo predmetnim oznakam: ožje ključne besede (Descriptors, Heading Words, Controlled Terms, Keywords) in širše predmetne kategorije (Subject Categories, Concept Codes, CABI Codes, Classification Codes). Podani so seštevki besed oz. indeksnih gesel ali pojmov z najvišjo pojavnostjo. S tem pregledom želimo spodbuditi rabo dodatnih podatkovnih zbirk ter rabo strukturiranih besednjakov, kar naj se izrazi v boljših iskalnih rezultatih. End-users should be able to identify the most relevant scientific and technical information resources. They must understand database structure, query techniques, controlled dictionaries, terminology and indexing schemes. In the present article, we review principal characteristics of the following information systems: bibliographic databases (Agricola, Agris, Cab Abstracts (CABI), Compendex, Inspec, Iconda, statistical collections Eurostat and Faostat, and technical standards and patents ISO and WIPO with regard to the field of wood sciences or processing of forest products. Differences in database fields and search characteristics (search platforms, Boolean logic, prefixes, suffixes, wildcards) are described. Special attention is placed on subject headings in each respective database: narrow keywords (Descriptors, Heading Words, Controlled Terms, Keywords) and broader categories (Subject Categories, Concept Codes, CABI Codes, Classification Codes). Word counts are shown for the terms with the highest frequency of occurrences. This review promotes the use of additional databases, encourages employment of structured vocabularies, which should result in improved search results.