X-Git-Url: http://git.rot13.org/?a=blobdiff_plain;f=README;h=55a3abb4e3d872b8439908a65b20d66082035bea;hb=fc69051198ec2102bfa800c7af64bb43e1ad33c4;hp=b2da9d11362572690c8a1010fd311a568139900b;hpb=df6d068ffdddf496ce775f4b0417743da886c39e;p=MojoFacets.git

diff --git a/README b/README
index b2da9d1..55a3abb 100644
--- a/README
+++ b/README
@@ -12,37 +12,51 @@ Multi-line values should be wrapped in ^multi-line-text^
 If you save bounch of html files with table in directory with .html extension, they will
 we all read as single data set, allowing easy analysys of search results, for example.
 
+CSV files with .csv extension are parsed using , as delimiter. Encoding is utf8
+and can be specified in filename, before extension like this: data.encoding.csv
+
+CouchDB data can be imported using files which contain full url to CouchDB database
+or url to CouchDB view to import. URL's filename should end in *.couchdb
+
 
 Start with:
 
     LANG=hr_HR.utf8 ./script/mojo_facets daemon --reload
 
 
-Data replication:
+Changing tabular data:
 
-    # master
-    ./script/mojo_facets daemon --listen 'http://*:4444' --reload
+Just double click on any table cell and click outside or focus out to save change.
 
-    # slave
-    MASTER=http://localhost:4444 ./script/mojo_facets daemon --reload
 
-
-Turning actions into changes:
+Data action and changes:
 
 There are two kinds of audit log in MojoFacets:
 
-1. actions stored in /tmp/actions are clicks on user interface with parameters
+1. actions stored in /tmp/actions are clicks on user interface with parameters,
+   they will probably be erased on next reboot since they are in /tmp
+
+2. changes in data/database.changes are more structured, including old value
+   and unique identifiers for that row
 
-2. changes are more structured, including old value and unique values from
-   edit of one value in dataset
+Changes can be applied on any dataset currently in memory.
 
-If you want to create changes which can be applied again on original dataset or
-any other dataset which has same unique values you can use helper script:
+Whole idea of changes is to create audit log which is detailed enough to recostruct
+current state of dataset from source file and list of changes. However, to speed
+up operations, you can periodically save your in-memory data to /tmp/ in perl
+storeable format using save link in interface.
 
-    $ ./script/actions-to-changes /data/mojo_facets/*
 
+Data replication:
 
-Modify your data using perl snippets:
+Actions can be replicated to other hosts using MASTER enviroment variable
+or config menu
+
+    # slave
+    MASTER=http://localhost:4444 ./script/mojo_facets daemon --reload
+
+
+Code console to modify data using perl snippets:
 
 Experimental REPL console supports perl snippets which get $row hash which is one
 element from your dataset.
@@ -50,6 +64,10 @@ element from your dataset.
 If you want to create or update values, you will have to use $update hash to set
 new values.
 
+If you want to report something from your dataset (also called reduce in map/reduce
+terminology) you can use $out hash to store values which will be used to generate
+new dataset using $key and $value for column names.
+
 All values are repetable, but if you create just a scalar, magic(tm) inside MojoFacets
 will try to upgrade it to [ $scalar ] so you don't have to do it explicitly.
 
@@ -59,6 +77,12 @@ They use column1,column2.description.pl notatition so only snippets which have a
 column will be shown.
 
 
+Facet code eval:
+
+Code snippet will be executed for each facet $value and will report $count and $checked state.
+You can also update $checked to programatically select part of facet values.
+
+
 Export data:
 
 All exported data is stored in public/export/database/