| Web site construction |
| The entire Spectra Online web site is hosted on a single Windows 2000 server. The server is running a number of Microsoft enterprise tools including Internet Information Server (IIS), Transaction Server (MTS) and SQL Server 7 database. |
| The operational parts of the web site (database search, data submission, user registration, etc.) were developed as Active Server Pages (ASP). These pages make extensive use of Spectral Server components to interact with the database, retrieve and display the spectra. |
| Compound, CAS, Formula, & Weight search |
| Depending on what type of search is selected, the Spectra Online web site dynamically builds a SQL query on specific columns in the database compound information. However, for compound searches, the query is always run on both the compound names and synonyms to ensure the best possible coverage of the compounds in the database. The ASP search code is designed to return the results list with a maximum of 100 compounds for any search to prevent overloading the server when a query is too broad. |
| Spectrum match search |
| Spectrum matching is one of the most powerful features of the Spectra Online web site. There are a number of steps involved to allow this operation to take place. First, there is the data file upload: the page allows users to upload a single data file in any data format for searching against the database. When the submitted file arrives on the web server, it is automatically recognized and translated into a form that can be compared to the spectral libraries using Thermo Electron's SmartConvert technology. The server also uses another Spectral Server component to automatically identify the type of spectrum (i.e., FT-IR, UV-Vis, MS, etc.). The translated spectrum is then rendered to a GIF file and dynamically combined with the determined spectrum information on a "preview" page which is sent back to the user's browser. |
| When the user clicks the button on the preview to initiate the search, the spectrum is digitally compared to every spectrum in a library containing only spectra of the same type (using yet another Spectral Server component). Since the spectral libraries are compressed representations of the full spectral data, spectral searches are very quick. Once the best set of matches is returned from the compressed library, the server looks up the matching records in the SQL database and returns the result list. In this case, the actual matching spectrum along with the Hit Quality Index (HQI) is shown in addition to the compound information. |
| Spectrum detail display |
| In order to display the details of a spectrum in the Spectra Online database, the web site offers users two ways to view the data: as interactive GIF images or using the Thermo Electron JavaView display applet. |
| The GIF imaging technology (used in the default display mode) was developed to support users on low-bandwidth connections such as dial up modems and slower WANs. Rather that transmit the entire spectrum data file (which can be very large) across the web, the Spectral Server display component automatically generates a small GIF image of the data. This is then embedded in a web page, which is sent back to the user's browser. The GIF image is set up as a mapped link to a script on the server that can detect where a user clicks on the image. Combined with the ability to read the settings in the "Click to" controls, this allows the server to render a new GIF image of an expanded area and update the web page. |
| Using the JavaView applet to display the spectrum is more substantially interactive than the GIF image as it allows users to select the zoom area using the classic box-pull method. However to use it, the browser must download the applet code and the full spectrum from the database. This was designed for users with higher-bandwidth connections such as fast WANs, corporate LANs, DSL, and cable modem Internet connections. |
| Another interesting feature on this part of the Spectra Online web site is the ability to download the spectrum from the database to the user's local computer. This is accomplished by pulling a copy of the spectral data from the database and sending it across the Internet as a binary file. The data is always transmitted in the Thermo Electron SPC file format and the server always identifies the data with the MIME type "chemical/x-galactic-spc". Any application on the user's local system that is set to handle that MIME type and can read SPC files can be used to view downloaded data files. For Spectra Online users running Windows 95/98/NT, there is a FREE viewer program for SPC data files available for download from Thermo Electron's web site. |
| Data submission |
| Unlike most spectral databases on the web, Spectra Online allows users to add new spectra to the database directly from a web browser. This is accomplished by following a series of steps each using different parts of the Spectral Server technology. |
| The first step is to upload the data file from the user's computer onto the Spectra Online server. Depending on the type of data file being submitted, one of two upload methods is employed. For data systems that store all the spectral data and other associated information in a single file, the standard upload form is used. This form just sends a copy of the user-selected data file up to the server. |
| However, not all instrument data systems store all their data in one file. For example, most NMR spectrometers store the data point values in one file and the instrument acquisition and processing parameters in separate files. In order for the data to be properly translated and stored in the database, all of these files must be sent to the server together. To address this, Thermo Electron developed a special "Web Data Posting" Java applet. When initiated, the applet downloads a list of instrument data formats that store information in multiple files from the Spectra Online server. When the user selects a file on the local system, the applet checks it against the list of data types and locates the additional files needed on the user's computer. When the user submits the data, all the files are automatically packed into a single ZIP file and sent to the server. |
| When the data arrives on the server (either as a single data file or multiple files packaged in ZIP file), they are translated into the database using the Spectral Server SmartConvert module. This is the very same module used when data is submitted for a Spectrum Match search of the database. The server then uses another Spectral Server component to automatically identify the type of spectrum (i.e., FT-IR, UV-Vis, MS, etc.). The translated spectrum is then rendered to a GIF file and dynamically combined with the determined spectrum information on a "preview" page which is sent back to the user's browser for approval. |
| Once the data is translated, the user must identify the compound used to measure the spectrum. If the compound is already in the Spectra Online database, the user can "link" the spectrum to that compound. (The spectral data and compound information are stored in separate tables in the database so that there can be any number of spectra attached to a given compound.) To accomplish this task, the user fills out a form to search the database by compound name or CAS Registry number. Similar to a compound lookup search, this creates a SQL query into the compound table and displays a list of matching compounds to link to the submitted spectrum. |
| If the compound does not exist, in the Spectra Online database, the user can create a new compound. When creating a new compound, the user fills in the relevant information (i.e., Compound Name, CAS Registry number, Molecular Formula, etc.) and submits it as well. This automatically adds a new entry into the compound table in the database and links the previously submitted spectrum to that record. |
| Although the data is instantly added to the Spectra Online database, Thermo Electron periodically reviews all user-submitted data. The reviewers do not check the data for consistency or validity, but instead look for inappropriate content in the database. Any data submitted with textual information that is not directly related to the spectrum or chemical substance or is deemed inappropriate for a publicly visited web site is immediately deleted. |
