Changing the Administration Password

1. On Windows

   First, make sure Flax Basic is not running. If you installed Flax Basic as a Service, you can stop it by running the file

   C:\Program Files\Flax Basic\stopflaxservice.bat

   or if it is running in a command window then close that window. Then open a command window (Start, Run, cmd) and type:

   cd "C:\Program files\Flax Basic" startflax --set-admin-password

   Enter a new password and close the window, then restart the Flax Basic service if necessary by running the file

   C:\Program Files\Flax Basic\startflaxservice.bat

2. Linux/Unix/MacOS

   First stop the Flax process, if it is running. Then enter the following on the command line: (FIXME)

   $ python startflax.py -d data --conf-dir=. --set-admin-password

   You will then be prompted to enter and confirm a new password. After entering it, restart Flax as ususal.

Creating a Collection

The Collections page is accessed by clicking the Collections link in the menu bar at the top of the page. When you run Flax for the first time, the collections list will be empty.

To create a new collection, click the Create a new collection link. This takes you to the collection details page, with blank fields. The fields are arranged in four sections:

1. Document Collection

   This section has two fields, Name and Description. The first is a short name which is used by Flax as a label for the collection, and is required. The description is optional, and can be of any length.

2. Files Specification

   This section controls which files will be included in the collection. Currently, Flax can only index files on file systems attached to the computer running Flax, including network file systems/shared folders (other access methods will be added in a later release of Flax.)

   Each collection requires one or more Path fields, which specifies the parent directory which will be scanned for files. Subdirectories will also be scanned recursively. If you want to specify more than one path, click the Add another path button. Path fields can be left blank to be omitted from the collection, so long as at least one has a value.

   Each Path has a corresponding Mapping field, which is used to generate URLs for search result objects. Files are not served directly through Flax - mostly as this could be used to circumvent any access restrictions implemented through the file system or other means. Any value in the mapping field will replace the path prefix in the generated URL.

   Flax is generally intended to work alongside an existing web server, ideally using it as a proxy (see above), and using it to serve documents from Flax's search results page. For example, using the main web server as a proxy, if documents from the folder C:\foo\bar\recipes is served as the URL /main/recipes, then you should enter:

   Path:     C:\foo\bar\recipes Mapping:  /main/recipes

   If the main web server is not being used as a proxy for Flax, you should enter a fully-qualified URL which specifies the server, e.g.:

   Path:     C:\foo\bar\recipes Mapping:  http://www.netveggie.foo/main/recipes

   You can also serve indexed files directly from Flax, by entering "FLAX" (upper-case) into the Mapping field (for each Path if there is more than one.) This is not likely to be as efficient as a dedicated web server. Also note that Flax will ignore any user access restrictions to files. For these reasons, we advise against using this function for a public web site or in a security-critical application.

   The Formats list allows you to choose file types to be included in the index. If you leave all unselected, Flax will index all file types.

   Finally, the Age field lets you exclude files with a modification time greater than the period selected. When re-indexing a collection, any files that now exceed the Age field (if set) will be removed.

3. Indexing Options

   Options controlling the way files are indexed. Currently, this is just a language selection, which controls the stemming, or suffix-stripping, of indexed words. This technique improves searching by normalising inflected forms of words, so that, for example, "cycle" would match "cycled", "cycling", "cycles" etc. Since this is a language-dependent feature, the main language of a collection should be chosen if stemming is to be used.

   If you have documents in several different languages, you can either index and search them as separate collections, with the language field set accordingly, or index them all with language set to None, in which case stemming will not be performed.

4. Scheduling

   Indexing can be initiated manually (see below) or at automatically at scheduled times (the Flax service must be running for this to work.) Scheduling works by matching the current time, once per minute, against the scheduling specification of a collection (if set). If all of the fields (Minutes, Hours, Weekdays, Monthdays and Months) matches the current time then the collection is scheduled for indexing. A blank field will never match, ensuring that no scheduled indexing will be performed if any field is blank. The "*" wildcard matches any value. Otherwise, the scheduler expects a comma-delimeted list of integers.

   For example, to schedule reindexing on the hour, every day, enter:

   Minutes:	0
   Hours:	*
   Weekdays:	*
   Monthdays:	*
   Months:	*

   To reindex at 3.30am every weekday (Mondays are 0, Sundays are 6):

   Minutes:	30
   Hours:	3
   Weekdays:	0, 1, 2, 3, 4
   Monthdays:	*
   Months:	*

After entering details for the new collection, click the Apply button at the bottom of the page. Flax will create the new collection and return you to the collections list.

Indexing a Collection

Before searching a collection, it must be indexed. This scans all the documents defined by the collection, extracts text from them, and creates a index on disk for optimised searching. Flax will only index one collection at a time, as this is the most efficient use of machine resources. More than one collection may be due for indexing, in which case they will be indexed in an arbitrary order. This may be controlled, where necessary, with the manual controls described below.

1. Unscheduled Indexing

   To index a collection immediately, assuming no other collections are being indexed, click the Start button in the Indexing column of the collections list. The button text will change to In progress and the information columns will update as the indexing progresses:

   Docs:	is the number of indexed documents in the collection.
   Files:	is the number of files processed on the indexing run.
   Errors:	is the number of files that failed to be indexed due to errors.

   When the indexing completes, the button text will change back to Start.

2. Scheduled Indexing

   Scheduled indexing is defined with the Collection Details fields described above. When a scheduled time arrives, indexing will become due for the collection, and it will be indexed when the indexer is free (or immediately if no other indexing is taking place). The collections list display will update in the same way as for unscheduled indexing.

3. Holding Indexing

   It is possible to hold a collection, which means that indexing is not started when it becomes due, either scheduled or unscheduled. To do this, click the Hold button in the collections list. The button text will change to Unhold, allowing the collection hold to be released when you decide. If a collection is already being indexed when held, it will be re-indexed from the start when it becomes unheld.

   Holding a collection could be useful when making significant changes to a set of documents (such as adding a large number from an archive) as it means indexing can be postponed until all changes are made. It also allows you to immediately index a collection even if other collections are currently being indexed - just put the other collections on hold.

Editing and Deleting a Collection

Any collection details (description, file specifications, indexing options, scheduling - everything apart from name) may be changed by clicking the collection name in the collections list. This opens the Collection Details page, where changes may be made and saved by clicking the Apply button. Changes to the file specifications and indexing options will not be reflected in the collection index until it is re-indexed.

To permanently delete a collection, click the delete icon (a red X) in the collections list. You will be asked to confirm before Flax deletes the collection.

Spelling correction

If a search returns no results, Flax will attempt to correct the spelling of any words in the query that appear to be misspelled, by searching the index for the nearest match to the word. If it finds a query that returns results, it will display these results along with a warning message that the query has been spell-corrected.