Commands processed by the Mirage interpreter

Syntax Summary

  • load dataFile [format formatFile] [skip comment] [[first #lines]|[random fraction[seed ranseed]]]
  • httpget serverName outputFile sqlCommand
  • addfrom dataFile
  • replacefrom dataFile
  • addcolfrom dataFile newAttr1 newAttr2 ...
  • addcol attr1 + attr2
  • addcol attr1 - attr2
  • addcol attr1 * attr2
  • select attr1 lower1 upper1 attr2 lower2 upper2 ...
  • or attr1 lower1 upper1 attr2 lower2 upper2 ...
  • and attr1 lower1 upper1 attr2 lower2 upper2 ...
  • clearselect
  • colorselect r g b
  • colorall r g b
  • pause milliSeconds
  • run kmeans vector vectorName nclus number
  • run hclus vector vectorName
  • run phclus vector vectorName nclus number
  • dbconn dbURL dbDriverClassName
  • dbquery SQLquery
  • dbclose
  • import formatStatement
  • source script
  • savejob
  • exit
  • Usage

    (each command should be on a single, new line)
    command function defaults
    load dataFile read a specific data file no default file
    load dataFile format formatFile read a specific data file and use format statements from a different file use format lines inside data file, or use the default format if there are no format lines
    load dataFile skip commentString read specified data file but skip all lines beginning with the comment string comment string = #
    load dataFile first 100 read the first 100 data lines of the specified data file (#lines > 0) read all lines
    load dataFile random 0.1 read a random fraction of specified data file (0.0 < fraction < 1.0) fraction = 1.0
    load dataFile random 0.1 seed 1 read a random fraction of specified data file (0.0 < fraction < 1.0) after changing the seed of the random number generator (seed > 0) fraction = 1.0; seed = 0
    httpget DLS dls.dat select ra,decl from photo where flags = 0 retrieve data via http from DLS server using a SQL query, and save the obtained data in a file dls.dat server need to be set up in Property.dat
    addfrom newDataFile add entries from a new data file assuming the same format no default file
    replacefrom newDataFile replace all entries by those in a new data file, assuming the same format. this will also remove all existing proximity structures, since they may not point to the correct data entries any more. no default file
    addcolfrom newDataFile add all the columns from a new data file optionally with format if no format, variables will be assigned default names
    addcol attr1 op attr2 add a column for the results of the operation (+,-,*) new column is named by the formula
    import formatStatement this adds a new format statement (without the "format" keyword) to an already loaded database. this may be the definition of a new vector, a modification of an existing vector's extent, inclusion of a new proximity structure, binding to a new image, etc.
    select attr1 lower1 upper1 attr2 lower2 upper2 ... select points with attribute attr1 between the given lower1 and upper1 limits (inclusive) and attribute attr2 between the respective limits, and so on. if the attribute is a text variable, the limits are interpreted as a desired prefix (lower limit) and suffix (upper limit) of the string. a wild card character * matches arbitrary prefixes and suffixes. any previous selection is cleared.
    or attr1 lower1 upper1 attr2 lower2 upper2 ... make a new selection with the given limits and join with the previous selection by a union operation.
    and attr1 lower1 upper1 attr2 lower2 upper2 ... make a new selection with the given limits and join with the previous selection by an intersection operation.
    clearselect clear any current selection.
    colorselect r g b paint current selection with color (r,g,b).
    colorall r g b paint all points with color (r,g,b).
    pause 1000 pause for 1000 milliseconds.
    run kmeans vector vectorName nclus 10 run kmeans clustering to make 10 clusters with a specific vector. vector must be defined in format already.
    run hclus vector vectorName run hierarchical clustering on a specific vector. vector must be defined in format already.
    run phclus vector vectorName nclus 10 run kmeans clustering to make 10 clusters with a specific vector, and then run hierarchical clustering on the cluster centers. vector must be defined in format already.
    dbconn dbURL dbDriverClassName open a connection to a database server at dbURL using driver dbDriverClassName (see JDBC documentation for details)
    dbquery query send an SQL query (a select statement) to a database server to which connection has been established by the dbconn command. convert the returned table to a Mirage dataset.
    dbclose close the existing connection to the database server
    source script execute commands contained in a specific script
    savejob save temporary job directories do not save
    exit terminate the interpreter