doc: adjust column width of PMD overview tables
[dpdk.git] / doc / guides / conf.py
index fba34a6..31f914a 100644 (file)
@@ -1,5 +1,5 @@
 #   BSD LICENSE
-#   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
+#   Copyright(c) 2010-2015 Intel Corporation. All rights reserved.
 #   All rights reserved.
 #
 #   Redistribution and use in source and binary forms, with or without
 #   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 #   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
+from __future__ import print_function
 import subprocess
+from docutils import nodes
+from distutils.version import LooseVersion
+from sphinx import __version__ as sphinx_version
+from sphinx.highlighting import PygmentsBridge
+from pygments.formatters.latex import LatexFormatter
+from os import listdir
+from os.path import basename
+from os.path import dirname
+from os.path import join as path_join
 
-project = 'DPDK'
+try:
+    # Python 2.
+    import ConfigParser as configparser
+except:
+    # Python 3.
+    import configparser
 
+try:
+    import sphinx_rtd_theme
+
+    html_theme = "sphinx_rtd_theme"
+    html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]
+except:
+    print('Install the sphinx ReadTheDocs theme for improved html documentation '
+          'layout: pip install sphinx_rtd_theme')
+    pass
+
+project = 'Data Plane Development Kit'
+html_logo = '../logo/DPDK_logo_vertical_rev_small.png'
+latex_logo = '../logo/DPDK_logo_horizontal_tag.png'
+html_add_permalinks = ""
 html_show_copyright = False
+highlight_language = 'none'
 
-version = subprocess.check_output(["make","-sRrC","../../", "showversion"])
+version = subprocess.check_output(['make', '-sRrC', '../../', 'showversion'])
+version = version.decode('utf-8').rstrip()
+release = version
 
 master_doc = 'index'
+
+# Maximum feature description string length
+feature_str_len = 25
+
+# Figures, tables and code-blocks automatically numbered if they have caption
+numfig = True
+
+latex_documents = [
+    ('index',
+     'doc.tex',
+     '',
+     '',
+     'manual')
+]
+
+# Latex directives to be included directly in the latex/pdf docs.
+custom_latex_preamble = r"""
+\usepackage[utf8]{inputenc}
+\usepackage[T1]{fontenc}
+\usepackage{helvet}
+\renewcommand{\familydefault}{\sfdefault}
+\RecustomVerbatimEnvironment{Verbatim}{Verbatim}{xleftmargin=5mm}
+"""
+
+# Configuration for the latex/pdf docs.
+latex_elements = {
+    'papersize': 'a4paper',
+    'pointsize': '11pt',
+    # remove blank pages
+    'classoptions': ',openany,oneside',
+    'babel': '\\usepackage[english]{babel}',
+    # customize Latex formatting
+    'preamble': custom_latex_preamble
+}
+
+
+# Override the default Latex formatter in order to modify the
+# code/verbatim blocks.
+class CustomLatexFormatter(LatexFormatter):
+    def __init__(self, **options):
+        super(CustomLatexFormatter, self).__init__(**options)
+        # Use the second smallest font size for code/verbatim blocks.
+        self.verboptions = r'formatcom=\footnotesize'
+
+# Replace the default latex formatter.
+PygmentsBridge.latex_formatter = CustomLatexFormatter
+
+# Configuration for man pages
+man_pages = [("testpmd_app_ug/run_app", "testpmd",
+              "tests for dpdk pmds", "", 1),
+             ("tools/pdump", "dpdk-pdump",
+              "enable packet capture on dpdk ports", "", 1),
+             ("tools/proc_info", "dpdk-procinfo",
+              "access dpdk port stats and memory info", "", 1),
+             ("tools/pmdinfo", "dpdk-pmdinfo",
+              "dump a PMDs hardware support info", "", 1),
+             ("tools/devbind", "dpdk-devbind",
+              "check device status and bind/unbind them from drivers", "", 8)]
+
+
+# ####### :numref: fallback ########
+# The following hook functions add some simple handling for the :numref:
+# directive for Sphinx versions prior to 1.3.1. The functions replace the
+# :numref: reference with a link to the target (for all Sphinx doc types).
+# It doesn't try to label figures/tables.
+def numref_role(reftype, rawtext, text, lineno, inliner):
+    """
+    Add a Sphinx role to handle numref references. Note, we can't convert
+    the link here because the doctree isn't build and the target information
+    isn't available.
+    """
+    # Add an identifier to distinguish numref from other references.
+    newnode = nodes.reference('',
+                              '',
+                              refuri='_local_numref_#%s' % text,
+                              internal=True)
+    return [newnode], []
+
+
+def process_numref(app, doctree, from_docname):
+    """
+    Process the numref nodes once the doctree has been built and prior to
+    writing the files. The processing involves replacing the numref with a
+    link plus text to indicate if it is a Figure or Table link.
+    """
+
+    # Iterate over the reference nodes in the doctree.
+    for node in doctree.traverse(nodes.reference):
+        target = node.get('refuri', '')
+
+        # Look for numref nodes.
+        if target.startswith('_local_numref_#'):
+            target = target.replace('_local_numref_#', '')
+
+            # Get the target label and link information from the Sphinx env.
+            data = app.builder.env.domains['std'].data
+            docname, label, _ = data['labels'].get(target, ('', '', ''))
+            relative_url = app.builder.get_relative_uri(from_docname, docname)
+
+            # Add a text label to the link.
+            if target.startswith('figure'):
+                caption = 'Figure'
+            elif target.startswith('table'):
+                caption = 'Table'
+            else:
+                caption = 'Link'
+
+            # New reference node with the updated link information.
+            newnode = nodes.reference('',
+                                      caption,
+                                      refuri='%s#%s' % (relative_url, label),
+                                      internal=True)
+            node.replace_self(newnode)
+
+
+def generate_overview_table(output_filename, table_id, section, table_name, title):
+    """
+    Function to generate the Overview Table from the ini files that define
+    the features for each driver.
+
+    The default features for the table and their order is defined by the
+    'default.ini' file.
+
+    """
+    # Default warning string.
+    warning = 'Warning generate_overview_table()'
+
+    # Get the default features and order from the 'default.ini' file.
+    ini_path = path_join(dirname(output_filename), 'features')
+    config = configparser.ConfigParser()
+    config.optionxform = str
+    config.read(path_join(ini_path, 'default.ini'))
+    default_features = config.items(section)
+
+    # Create a dict of the valid features to validate the other ini files.
+    valid_features = {}
+    max_feature_length = 0
+    for feature in default_features:
+        key = feature[0]
+        valid_features[key] = ' '
+        max_feature_length = max(max_feature_length, len(key))
+
+    # Get a list of driver ini files, excluding 'default.ini'.
+    ini_files = [basename(file) for file in listdir(ini_path)
+                 if file.endswith('.ini') and file != 'default.ini']
+    ini_files.sort()
+
+    # Build up a list of the table header names from the ini filenames.
+    header_names = []
+    for ini_filename in ini_files:
+        name = ini_filename[:-4]
+        name = name.replace('_vf', 'vf')
+
+        # Pad the table header names to match the existing format.
+        if '_vec' in name:
+            pmd, vec = name.split('_')
+            name = '{0:{fill}{align}7}vec'.format(pmd, fill='.', align='<')
+        else:
+            name = '{0:{fill}{align}10}'.format(name, fill=' ', align='<')
+
+        header_names.append(name)
+
+    # Create a dict of the defined features for each driver from the ini files.
+    ini_data = {}
+    for ini_filename in ini_files:
+        config = configparser.ConfigParser()
+        config.optionxform = str
+        config.read(path_join(ini_path, ini_filename))
+
+        # Initialize the dict with the default.ini value.
+        ini_data[ini_filename] = valid_features.copy()
+
+        # Check for a valid ini section.
+        if not config.has_section(section):
+            print("{}: File '{}' has no [{}] secton".format(warning,
+                                                            ini_filename,
+                                                            section))
+            continue
+
+        # Check for valid features names.
+        for name, value in config.items(section):
+            if name not in valid_features:
+                print("{}: Unknown feature '{}' in '{}'".format(warning,
+                                                                name,
+                                                                ini_filename))
+                continue
+
+            if value is not '':
+                # Get the first letter only.
+                ini_data[ini_filename][name] = value[0]
+
+    # Print out the RST Driver Overview table from the ini file data.
+    outfile = open(output_filename, 'w')
+    num_cols = len(header_names)
+
+    print_table_css(outfile, table_id)
+    print('.. table:: ' + table_name + '\n', file=outfile)
+    print_table_header(outfile, num_cols, header_names, title)
+    print_table_body(outfile, num_cols, ini_files, ini_data, default_features)
+
+
+def print_table_header(outfile, num_cols, header_names, title):
+    """ Print the RST table header. The header names are vertical. """
+    print_table_divider(outfile, num_cols)
+
+    line = ''
+    for name in header_names:
+        line += ' ' + name[0]
+
+    print_table_row(outfile, title, line)
+
+    for i in range(1, 10):
+        line = ''
+        for name in header_names:
+            line += ' ' + name[i]
+
+        print_table_row(outfile, '', line)
+
+    print_table_divider(outfile, num_cols)
+
+
+def print_table_body(outfile, num_cols, ini_files, ini_data, default_features):
+    """ Print out the body of the table. Each row is a NIC feature. """
+
+    for feature, _ in default_features:
+        line = ''
+
+        for ini_filename in ini_files:
+            line += ' ' + ini_data[ini_filename][feature]
+
+        print_table_row(outfile, feature, line)
+
+    print_table_divider(outfile, num_cols)
+
+
+def print_table_row(outfile, feature, line):
+    """ Print a single row of the table with fixed formatting. """
+    line = line.rstrip()
+    print('   {:<{}}{}'.format(feature, feature_str_len, line), file=outfile)
+
+
+def print_table_divider(outfile, num_cols):
+    """ Print the table divider line. """
+    line = ' '
+    column_dividers = ['='] * num_cols
+    line += ' '.join(column_dividers)
+
+    feature = '=' * feature_str_len
+
+    print_table_row(outfile, feature, line)
+
+
+def print_table_css(outfile, table_id):
+    template = """
+.. raw:: html
+
+   <style>
+      .wy-nav-content {
+         opacity: .99;
+      }
+      table#idx {
+         cursor: default;
+         overflow: hidden;
+      }
+      table#idx th, table#idx td {
+         text-align: center;
+      }
+      table#idx th {
+         font-size: 80%;
+         white-space: pre-wrap;
+         vertical-align: top;
+         padding: 0.5em 0;
+         min-width: 0.9em;
+         width: 2em;
+      }
+      table#idx col:first-child {
+         width: 0;
+      }
+      table#idx th:first-child {
+         vertical-align: bottom;
+      }
+      table#idx td {
+         font-size: 70%;
+         padding: 1px;
+      }
+      table#idx td:first-child {
+         padding-left: 1em;
+         text-align: left;
+      }
+      table#idx tr:nth-child(2n-1) td {
+         background-color: rgba(210, 210, 210, 0.2);
+      }
+      table#idx th:not(:first-child):hover,
+      table#idx td:not(:first-child):hover {
+         position: relative;
+      }
+      table#idx th:not(:first-child):hover::after,
+      table#idx td:not(:first-child):hover::after {
+         content: '';
+         height: 6000px;
+         top: -3000px;
+         width: 100%;
+         left: 0;
+         position: absolute;
+         z-index: -1;
+         background-color: #ffb;
+      }
+      table#idx tr:hover td {
+         background-color: #ffb;
+      }
+   </style>
+"""
+    print(template.replace("idx", "id%d" % (table_id)), file=outfile)
+
+
+def setup(app):
+    table_file = dirname(__file__) + '/nics/overview_table.txt'
+    generate_overview_table(table_file, 1,
+                            'Features',
+                            'Features availability in networking drivers',
+                            'Feature')
+    table_file = dirname(__file__) + '/cryptodevs/overview_feature_table.txt'
+    generate_overview_table(table_file, 1,
+                            'Features',
+                            'Features availability in crypto drivers',
+                            'Feature')
+    table_file = dirname(__file__) + '/cryptodevs/overview_cipher_table.txt'
+    generate_overview_table(table_file, 2,
+                            'Cipher',
+                            'Cipher algorithms in crypto drivers',
+                            'Cipher algorithm')
+    table_file = dirname(__file__) + '/cryptodevs/overview_auth_table.txt'
+    generate_overview_table(table_file, 3,
+                            'Auth',
+                            'Authentication algorithms in crypto drivers',
+                            'Authentication algorithm')
+    table_file = dirname(__file__) + '/cryptodevs/overview_aead_table.txt'
+    generate_overview_table(table_file, 4,
+                            'AEAD',
+                            'AEAD algorithms in crypto drivers',
+                            'AEAD algorithm')
+
+    if LooseVersion(sphinx_version) < LooseVersion('1.3.1'):
+        print('Upgrade sphinx to version >= 1.3.1 for '
+              'improved Figure/Table number handling.')
+        # Add a role to handle :numref: references.
+        app.add_role('numref', numref_role)
+        # Process the numref references once the doctree has been created.
+        app.connect('doctree-resolved', process_numref)