hatchet · slabasan · Jul 10, 2020 · Jul 7, 2020
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,3 +1,35 @@
+# v1.2.0 (2020-07-07)
+
+This release adds a syntax query language as an alternative method for
+filtering the graph. It also refreshes the tree printout with an updated
+format and legend.
+
+### New features
+
+* Add graph syntax query language to filter the graph
+* Update HPCToolkit reader to handle sec or usec time units
+
+### Changes to existing APIs
+
+* Add squash parameter to filter function to perform filter and squash in a
+  single call
+* Filter function takes an object, which can be a user-supplied function or a
+  query object
+* Tree printout format updated
+* Tree printout API parameter changes:
+    - Removed parameters: ``color`` deprecated (color determined based on terminal support),
+      ``threshold``, and ``unicode``
+    - Changed parameters: ``metric`` changed to ``metric_column``, ``name`` changed to
+      ``name_column``, ``invert_colors`` changed to ``invert_colormap``,
+      ``expand_names`` changed to ``expand_name``, and ``context`` changed to
+      ``context_column``
+    - added ``highlight_name`` to highlight user code (from non-user code)
+
+### Bugfixes
+
+* Sort nodes in union and tree printout by their frame
+* Fix squash edge case where multiple parents are the same
+
 # v1.1.0 (2020-05-07)
 
 This release adds new analysis operators, as well as some bugfixes and minor
@@ -8,7 +40,7 @@ changes.
 * Add GraphFrame reindex operator
 * Query hatchet module version
 
-### Changes to existing APIs changes
+### Changes to existing APIs
 
 * Add depth parameter to tree printout
 

diff --git a/docs/advanced_examples.rst b/docs/advanced_examples.rst
@@ -147,9 +147,9 @@ biggest offenders.
   squashed_gf2 = filtered_gf2.squash()
 
   diff_gf = squashed_gf2 - squashed_gf1
- 
+
   sorted_df = diff_gf.dataframe.sort_values(by=['time'], ascending=False)
-  print(sorted_df) 
+  print(sorted_df)
 
 .. image:: images/lulesh-mpi.png
    :scale: 40 %

diff --git a/docs/images/query-dataframe.png b/docs/images/query-dataframe.png
diff --git a/docs/user_guide.rst b/docs/user_guide.rst
@@ -91,7 +91,7 @@ When the graph represented by the input dataset is small, the user may be intere
 
 .. code-block:: python
 
-  print(gf.tree(color=True))
+  print(gf.tree())
 
 One can also use the ``to_dot()`` function to output the tree as a string in the Graphviz' DOT format. This can be written to a file and then used to display a tree using the ``dot`` or ``neato`` program.
 
@@ -149,26 +149,49 @@ Dataframe operations
    :scale: 40 %
    :align: right
 
-**filter**: ``filter`` takes a user-supplied function and applies that to all
-rows in the DataFrame. The resulting Series or DataFrame is used to filter the
-DataFrame to only return rows that are true. The returned GraphFrame preserves
-the original graph provided as input to the filter operation. The images on the
-right show a DataFrame before and after a filter operation.
+**filter**: ``filter`` takes a user-supplied function or query object and
+applies that to all rows in the DataFrame. The resulting Series or DataFrame is
+used to filter the DataFrame to only return rows that are tru
8000
e. The returned
+GraphFrame preserves the original graph provided as input to the filter
+operation.
 
 .. code-block:: python
 
   filtered_gf = gf.filter(lambda x: x['time'] > 10.0)
 
+The images on the right show a DataFrame before and after a filter
+operation.
+
 .. image:: images/filter-dataframe.png
    :scale: 40 %
    :align: right
 
+An alternative way to filter the DataFrame is to supply a query object. The
+example below looks for a path that matches first a single node with name
+`solvers`, followed by 0 or more nodes with an inclusive time greater than 10,
+followed by a single node with name that starts with `p` and ends in an integer
+and has an inclusive time greater than or equal to 10.
+
 Filter is one of the operations that leads to the graph object and DataFrame
 object becoming inconsistent. After a filter operation, there are nodes in the
 graph that do not return any rows when used to index into the DataFrame.
 Typically, the user will perform a squash on the GraphFrame after a filter
 operation to make the graph and DataFrame objects consistent again.
 
+.. image:: images/query-dataframe.png
+   :scale: 35 %
+   :align: right
+
+.. code-block:: python
+
+  query = [
+      {"name": "solvers"},
+      ("*", {"time (inc)": "> 10"}),
+      {"name": "p[a-z]+[0-9]", "time (inc)": ">= 10"}
+  ]
+
+  filtered_gf = gf.filter(query)
+
 **drop_index_levels**: When there is per-MPI process or per-thread
 data in the DataFrame, a user might be interested in aggregating the data in
 some fashion to analyze the graph at a coarser granularity. This function

diff --git a/hatchet/__init__.py b/hatchet/__init__.py
@@ -9,5 +9,5 @@
 from .graphframe import GraphFrame
 from .query_matcher import QueryMatcher
 
-__version_info__ = ("1", "1", "0")
+__version_info__ = ("1", "2", "0")
 __version__ = ".".join(__version_info__)
diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "hatchet"
-version = "1.1.0"
+version = "1.2.0"
 description = "A Python library for analyzing hierarchical performance data."
 authors = [
     "Abhinav Bhatele <bhatele@cs.umd.edu>",