Sign on

Welcome to the ADS LANL Preprints Abstract Service!

This service provides access to six months worth of preprints which have been submitted to astronomical journals and conference proceedings. The data are provided by the preprint archive at Los Alamos National Laboratory. We have included all preprints in their "astro-ph" group, including any which have cross links to other groups. Preprints in our system expire after 6 months because on average, preprints which have been accepted to the major journals are published within that 6 month timeframe, and are therefore available in our journal Abstract Service.

Users can submit queries by author, date range, words in the title, and words in the abstract text. The resulting list is ranked by how closely the paper matched the query, and from this list, one can obtain all of the information contained in the ADS Preprint database (including the bibliographic code, title, authors, journal information, publication date, origin, and abstract text).

This document contains the following:

About the Data
The Abstract Query Form
Filters
Changing the Settings
Settings Example
Server Attributes
Table of Contents Query
Results List Window
Available Items
Find Similar Abstracts
What's New Service
Give us Feedback!
More Information about ADS
If you have questions about the Abstract Service, please send an e-mail message to ads@cfa.harvard.edu.

About the Data

The preprint database contains the most recent six months of astro-ph data from the LANL preprint archive. Users wishing to submit data should do so through their site.

The Abstract Query Form

The Abstract Query Form provides basic access into the ADS LANL preprint database. The query form is divided into 3 sections: The main search parameters, the filters, and the settings. To create a query, fill in any or all of the following fields: (ALL FIELDS ARE CASE-INSENSITIVE!).

Authors
Authors may be entered one per line, or on the same line separated by semi-colons. Authors may be entered by last name only, or by last name and first initial. Any of the following will work:
              smith, a.
              smith, a
              smith,a.
              smith,a
If a full first name is entered, only the first initial will be used by default. To search for middle initials of authors, use the Exact Author search capability.. The list of authors may be searched by doing a List Query.

Note that authors whose last names contain an umlaut may be entered under both possible English spellings (e.g. Boehm and Bohm).

Date
The "Publication Date" should be entered as two integers in the form MM and YYYY (e.g. 12 and 1988). For those cases when there was no month available, the month is entered in the database as 00.

If no date is entered, the program will default to the full date range. If no "From" month is entered the program will enter a default value of 00. If no "To" month is entered the program will enter a default value of 12.

Title
This allows the user to search on words found in the title of the paper. A checkbox for boolean logic allows combinations the use of boolean logic constructs.

Abstract Text
This allows the user to enter any combination of words or sentences he or she chooses. The search will be done on individual words (except for some of the most common English words such as "the", "a", and "and"). This is extremely useful for entering the text of an abstract from a previous query to find all the papers most related to that paper. Note, all title words and keywords are also indexed together with the abstract text words. A checkbox for boolean logic allows combinations the use of boolean logic constructs.

Filters

The filters sections allows the user to further modify the search.

Entry Date
The "Entry Date" allows the user to select data which has been entered in the database since a given date. Entering "-31" in the "Day" field will select all new entries in the past month.

Minimum Score
The "Minimum Score" allows the user to select data which has a score greater than the entered minimum score. This is most useful when used in conjunction with the entry date. See the "What's New Service" described below.

To change the logic within a field or between fields, you need to Change the Settings.

Changing the Settings

The settings section allows the user to change default query conditions such as the logic of the query. The following may be changed:

Relative Weights
This is the relative weighting of the fields used in calculating the scores. If you want to give more weight to the authors than the keywords, for example, these numbers would be changed. It is also possible to make one of these fields 0.0, in which case that item will not be considered when the relative scores of each abstract are calculated (the "scores" are descriptions of how well a given abstract matches the query conditions, normalized to 1.0).

SYNONYM REPLACEMENT
If this is turned on, the standard term list will be used to replace words. This corrects for misspellings in the abstract text or author names, and equates words such as "accelerate", "accelerated", and "acceleration".

It is also possible to turn synonym replacement on or off for individual words within a query. By default, synonym replacement is done for all words. To exclude a word from synonym replacement, use the "=" sign before that word (to exactly match that word and no synonyms). If you have turned off synonym replacement but want it on for a given word, use the "#" sign before the word.

SIMPLE LOGIC QUERIES
The simple logic recognizes '+' and '-' before the words. To require a word to be found in a search, it needs a '+' in front of it. A '-' before a word indicates that only references that do NOT contain that word should be returned. AND , OR, NOT are stopwords and will be ignored in the simple logic.

Note that if you want to exclude a word, it is best if it is not the first or only word listed in the search. This requires every abstract in the system to be returned, before the excluded word can be found and omitted from the returned abstracts.

It is also possible to turn synonym replacement on or off for individual words within a query. By default, synonym replacement is done for all words. To exclude a word from synonym replacement, use the "=" sign before that word (to exactly match that word and no synonyms). If you have turned off synonym replacement but want it on for a given word, use the "#" sign before the word.

FULL BOOLEAN QUERIES
This allows more complex queries than just combining all words with "OR" or "AND". The allowed boolean operators are: "and", "or", "not", "(", and ")". They can be used in any combination (as long as "(" and ")" match). For example the query "(redshift or survey) and not galaxy" finds all references that contain either "redshift" or "survey", but not "galaxy". The order of precedence of the operators is "(...)", "not", "and", "or". For larger numbers of search terms, this type of search will be slower than regular searches, especially if the "not" operator is used. For full boolean author searches, the syntax is somewhat awkward. Each operator has to be on a separate line. This includes the "(" and ") operators. Regular scoring is done on any terms that are combined with "or". "and" and "not" combinations are scored as 1.

USE FOR WEIGHTING?
This determines if the field should be used in calculating the scores. If this is turned on, then this search field (authors, keywords, etc) will be included in the calculation for the total score of each abstract. If it is not turned on, the search field will be used for selection as specified in the other flags but will be ignored for the score calculation. This serves two main functions, to selectively turn off portions of a complex query without editing the query, and in conjunction with the REQUIRE FIELD FOR SELECTION flag to permit a Boolean search on a field to be combined with a relevancy ranking.

WEIGHTED SCORING
This sets how the score for each abstract is calculated. If this is not turned on, the scoring is "straight"; each query condition that is met (a "hit", e.g. if the abstract contains a keyword specified in the query) receives a score of 1 (before normalization). If this is turned on, the scoring is "weighted"; the score for a "hit" is the inverse log of the frequency of the specified condition (e.g. keyword) in the total database. This gives a higher score to "hits" of rarely used words since they are presumably a stronger search criterion.

REQUIRE FIELD FOR SELECTION
This determines how the words across fields are combined. If this is turned on, an abstract will only fulfill a query when the given field conditions are met. For example, if you give an author and set this to "yes" in the author Query Settings, all abstracts which are retrieved must contain that author. When this is not turned on, abstracts which do not contain that author, but which meet other conditions in the query will also be returned.

When the query has been set up to your satisfaction, press Send to send the query.

When the query has finished, a ranked list of papers that fulfilled the query requirements is returned. To view more information on a given paper, click on the bibliographic code. Or select the checkbox for those papers of interest and use the "Retrieve Selected Abstracts" button at the bottom of the page.

If you would like to see more than the first 50 references, change the appropriate numbers on the query form.

Settings Example

Assume there is a list of desired authors in the author field in the Abstract Query window. Assume that one wants papers where at least one of the desired names is an author, but does not want the paper to appear more relevant if more than one desired author is a co-author. In the Author Query Settings window, you would turn off USE FOR WEIGHTING, turn off REQUIRE FIELD FOR SELECTION, and turn on REQUIRE FIELD FOR SELECTION which would allow the rest of the query to determine the relevancy score.

Server Attributes

The abstract server has a number of features which have been developed in order to improve searching the dataset. These include the use of synonyms, stop words, and word translations.

By default, each search term specified in the title, abstract, or author query field triggers a search for all the term's synonyms as well as the term itself. The use of synonym searches for text words has been established to equate different tenses of the same verb, different genders of the same name, as well as different words with the same meaning. In addition, we use synonyms to perform foreign language translations, to equate foreign words with their English counterpart. This feature can be turned off by modifying the default settings. For authors, synonym replacement allows the user to search on an author's name and get back any alternative or incorrect spellings of which we are aware. For example, the author synonymn pair "A'HEARN, M" and "AHEARN, M" allows users to search on either spelling and retrieve abstracts which contain either spelling. If you are aware of an author name, misspelling, or foreign word translation which does not seem to be working correctly, please inform us by email to ads@cfa.harvard.edu

We maintain a list of stop words which are words that are removed from queries before they are processed. These are words commonly used in the English language with great frequency, as well as adverbs, prepositions and any other words not carrying a significant meaning when used in a scientific context.

Last, we perform word translations for certain terms or patterns which are conventionally written in a few different ways. This is most commonly done with astronomical object names (e.g. "M 31" and "M31"), as well as some composite words (e.g. "X RAY", "X-RAY" and "XRAY"). The list of words which are currently translated is available. If you are aware of others which should be included, please inform us by email to ads@cfa.harvard.edu

Table of Contents Queries

To obtain a list of articles which appear in a major journal, you can do a Table of Contents Query. Select the year, month or volume number for the desired journal. A Table of Contents query generates a list of references for the papers available from the specified journal.

Note that we have complete table of contents only for those journals which we are scanning and putting full journal articles online (ApJL, ApJ, AJ, PASP). For these journals, the month filter will work as expected. For the other journals, a toc query will return only what is in our database (not necessarily complete), and the month filter will not work as expected.

To obtain a list of articles which appear in a minor journal or conference proceedings, you need to obtain the partial bibliographic code from the list of journals and do a Bibliographic Code Query.

Results List Window

The Query Results window contains a list of all abstracts which fulfilled the input query. This list contains the bibliographic code, score, publication date, authors, title, and list of available items. To save or print this summary list, use the features provided in the File menu of your browser, or use the options at the bottom of the screen.

To retrieve the rest of the abstract information (including journal reference, author affiliations, keywords, and abstract text), you have one of the following options:

The selected references can be returned in a variety of formats: Please note that the BibTeX references make user of standard macros for identifying the popular journals.

Available Items

This feature shows at a glance what items are available for a particular abstract. This allows the user to see, for example, which articles have abstracts or full text articles available. The following abbreviations are used:

Find Similar Abstracts

Through this feature, the ADS Abstract Service provides the capability of finding all relevant abstracts in the database. This is done by modifying the query based on the results of a previous query. This feature is only available for individually retrieved abstracts.

WARNING: A very long query can take several minutes to return. One can speed the query up substantially by limiting the date range in the Query Form.

What's New Service

We have enhanced the ADS Abstract Service with the capability of searching the database for recent entries. This service allows the user to conduct searches of the abstract database for all new abstracts which fulfill his or her favorite query. The service can be used through the WWW or through email, and can be automated to return all new entries within a specified period (such as the last month).

To use this service:

1. Build the query:

Use the WWW interface to build a query by repeatedly executing it until the desired result is returned. If you do not have WWW access, you can request the query form by email to "adsquery@cfa.harvard.edu". A message body containing "help" will provide some assistance. To set up a query, the message body should include:

     action=http://adsabs.harvard.edu/abstract_service.html
     method=get
     return=form
This returns the form with hyperlinks surrounded by "@" signs. To follow a hyperlink, uncomment the appropriate one in the section which begins "Start of hyperlinks". To submit the form, fill out the form in the section which begins "Begin of Form 1" by filling in search criteria after the "=" signs.

2. Save the query form:

Request that the query form be returned for this query by selecting "Return Form" in the Build Query Form section below the send button. A negative number in the "Entry Date" day field determines how far back the query goes. For example, to get any entries in the last month, use -31.

3. Use the form:

This form can be used either through the WWW by opening it in the browser (or saving it in your hotlist) with the syntax:

     file:/directory/filename
or it can be used to execute an e-mail query by typing:

     mail adsquery@cfa.harvard.edu < /directory/filename
If you are on a Unix machine, you can set up a crontab file which will automatically send this email message at specified intervals. To enable this feature, edit your crontab file by issuing the command "crontab -e" and entering the following:

  0 24 1 * * mail adsquery@cfa.harvard.edu < /directory/filename >/dev/null
where /directory/filename is the name of your saved query file.

This will execute the query on the first of every month and query back for 31 days (if the day field contains -31).

Give us Feedback!

You can help improve our service! If you have comments for us, please submit them through our feedback form. We always welcome comments and suggestions for improvements.

If you would like to submit a preprint abstract, please do so through the LANL preprint archive site.

In addition, you can supply author comments about papers for which you are an author. This comment will be linked to the abstract so that it will be displayed whenever the paper in question is retrieved by a user. This is a useful way to incorporate errata or make additional information available. The author comments should be submitted through our forms interface.

These comments can be plain text or fully formatted html documents. They can include comments about the abstract, clarifications or more extended descriptions, and links to other abstracts and documents. We reserve the right to edit or reject any author comments that are inappropriate. For an example of an author comment, see Mike Kurtz's PhD Thesis.


If you wish to acknowledge us in a publication, kindly use a phrase such as the following: "This research has made use of NASA's Astrophysics Data System Abstract Service." Thanks!

More Information about ADS

More information about the ADS project is available here, or you can send e-mail to us at ads@cfa.harvard.edu