Improve the documentation

kif · Sep 21, 2021 · 9e2e316 · 9e2e316
1 parent b3d917c
commit 9e2e316
Show file tree

Hide file tree

Showing 3 changed files with 25 additions and 8 deletions.
diff --git a/doc/source/coverage.rst b/doc/source/coverage.rst
@@ -0,0 +1,16 @@
+Test coverage report for dahu
+=============================
+
+Measured on *dahu* version 1.0.0, 21/09/2021
+
+.. csv-table:: Test suite coverage
+   :header: "Name", "Stmts", "Exec", "Cover"
+   :widths: 35, 8, 8, 8
+
+   "factory.py", "86", "29", "33.7 %"
+   "job.py", "330", "76", "23.0 %"
+   "plugin.py", "101", "51", "50.5 %"
+   "utils.py", "45", "4", "8.9 %"
+   "plugins/example.py", "35", "24", "68.6 %"
+
+   "dahu total", "597", "184", "30.8 %"
diff --git a/doc/source/dahu.rst b/doc/source/dahu.rst
@@ -6,20 +6,20 @@ Dahu: online data analysis server
 The *dahu* server executes **jobs**:
 ------------------------------------
 
-* Each job lives in its own thread (yes, thread, not process, it the plugin developper to ensure the work he is doing is GIL-compliant).
-* Each job executes one plugin, provided by the plugin developper (i.e. the scientist)
+* Each job lives in its own thread (yes, thread, not process, it the plugin's developper to ensure the work he is doing is GIL-compliant).
+* Each job executes one plugin, provided by the plugin's developper (i.e. the scientist)
 * The job de/serialises JSON strings coming from/returning to Tango
 * Jobs are executed asynchronously, the request for calculation is answered instantaneously with a *jobid*.
-* The *jobid* can be used to poll the server for the status of the job or for manual synchronization (Tango can time-out!).
+* The *jobid* can be used to poll the server for the status of the job or for manual synchronization (mind that Tango can time-out!).
 * When jobs are finished, the client is notified via Tango events about the status
 * Results can be retrieved after the job has finished.
 
 Jobs execute **plugin**:
 ------------------------
 
-* Plugins are written in Python
-* Plugins can be classes or simple function
-* The input and output must be JSON-seriablisable as simple dictionnaries
+* Plugins are written in Python (extension in Cython or OpenCL are common)
+* Plugins can be classes or simple functions
+* The input and output MUST be JSON-seriablisable as simple dictionnaries
 * Plugins are dynamically loaded from Python modules
 * Plugins can be profiled for performance analysis
 
@@ -30,12 +30,12 @@ All jobs can be run offline using the `dahu-reprocess` command line tool.
 This tool is not multithreaded and plugins are directly run, it is intended for:
 
 * offline developments
-* re-processing some failed online processing.
+* re-processing some failed online processing (where performances are less critical).
 
 Dahu is light !
 ---------------
 
 Dahu is a small project started at ESRF in 2013 with less than 1000 lines of code.
 It is used in production since then on a couple of beamlines.
-With its FIFO scheduler, dahu is very fast (1µs, 0.3ms from Tango)
+With its FIFO scheduler, `dahu` is very fast (1µs locally, 0.3ms from Tango)
 
diff --git a/doc/source/index.rst b/doc/source/index.rst
@@ -13,6 +13,7 @@ Contents:
 
    dahu
    installation
+   coverage
    api/modules
 
 Indices and tables