Refactor inventory code out of Runner.

This introduces the Inventory class.

Playbook uses the internals of Runner to limit the number of hosts to poll
asynchronously. To accomodate this, Inventory can be restricted to specific
hosts.

bin/ansible

@@ -98,11 +98,11 @@ class Cli(object):
 
     # ----------------------------------------------
 
-    def get_polling_runner(self, old_runner, hosts, jid):
+    def get_polling_runner(self, old_runner, jid):
         return ansible.runner.Runner(
             module_name='async_status', module_path=old_runner.module_path,
             module_args="jid=%s" % jid, remote_user=old_runner.remote_user,
-            remote_pass=old_runner.remote_pass, host_list=hosts,
+            remote_pass=old_runner.remote_pass, inventory=old_runner.inventory,
             timeout=old_runner.timeout, forks=old_runner.forks,
             remote_port=old_runner.remote_port, pattern='*', 
             callbacks=self.silent_callbacks, verbose=True,
@@ -138,8 +138,10 @@ class Cli(object):
 
             clock = options.seconds
             while (clock >= 0):
-                polling_runner = self.get_polling_runner(runner, poll_hosts, jid)
+                runner.inventory.restrict_to(poll_hosts)
+                polling_runner = self.get_polling_runner(runner, jid)
                 poll_results = polling_runner.run()
+                runner.inventory.lift_restrictions()
                 if poll_results is None:
                     break
                 for (host, host_result) in poll_results['contacted'].iteritems():

lib/ansible/inventory.py

@@ -0,0 +1,189 @@
+# (c) 2012, Michael DeHaan <michael.dehaan@gmail.com>
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+
+#############################################
+
+import fnmatch
+import os
+import subprocess
+
+import constants as C
+from ansible import errors
+from ansible import utils
+
+class Inventory(object):
+    """ Host inventory for ansible.
+
+    The inventory is either a simple text file with systems and [groups] of 
+    systems, or a script that will be called with --list or --host.
+    """
+
+    def __init__(self, host_list=C.DEFAULT_HOST_LIST, extra_vars=None):
+
+        self._restriction = None
+
+        if type(host_list) == list:
+            self.host_list = host_list
+            self.groups = dict(ungrouped=host_list)
+            self._is_script = False
+            return
+
+        inventory_file = os.path.expanduser(host_list)
+        if not os.path.exists(inventory_file):
+            raise errors.AnsibleFileNotFound("inventory file not found: %s" % host_list)
+
+        self.inventory_file = os.path.abspath(inventory_file)
+
+        if os.access(self.inventory_file, os.X_OK):
+            self.host_list, self.groups = self._parse_from_script(extra_vars)
+            self._is_script = True
+        else:
+            self.host_list, self.groups = self._parse_from_file()
+            self._is_script = False
+
+    # *****************************************************
+    # Public API
+
+    def list_hosts(self, pattern="all"):
+        """ Return a list of hosts [matching the pattern] """
+        if self._restriction is None:
+            host_list = self.host_list
+        else:
+            host_list = [ h for h in self.host_list if h in self._restriction ]
+        return [ h for h in host_list if self._matches(h, pattern) ]
+
+    def restrict_to(self, restriction):
+        """ Restrict list operations to the hosts given in restriction """
+        if type(restriction)!=list:
+            restriction = [ restriction ]
+
+        self._restriction = restriction
+
+    def lift_restriction(self):
+        """ Do not restrict list operations """
+        self._restriction = None
+
+    def get_variables(self, host, extra_vars=None):
+        """ Return the variables associated with this host. """
+
+        if not self._is_script:
+            return {}
+
+        return self._get_variables_from_script(host, extra_vars)
+
+    # *****************************************************
+
+    def _parse_from_file(self):
+        ''' parse a textual host file '''
+
+        results = []
+        groups = dict(ungrouped=[])
+        lines = file(self.inventory_file).read().split("\n")
+        group_name = 'ungrouped'
+        for item in lines:
+            item = item.lstrip().rstrip()
+            if item.startswith("#"):
+                # ignore commented out lines
+                pass
+            elif item.startswith("["):
+                # looks like a group
+                group_name = item.replace("[","").replace("]","").lstrip().rstrip()
+                groups[group_name] = []
+            elif item != "":
+                # looks like a regular host
+                groups[group_name].append(item)
+                if not item in results:
+                    results.append(item)
+        return (results, groups)
+
+    # *****************************************************
+
+    def _parse_from_script(self, extra_vars=None):
+        ''' evaluate a script that returns list of hosts by groups '''
+
+        results = []
+        groups = dict(ungrouped=[])
+
+        cmd = [self.inventory_file, '--list']
+
+        if extra_vars:
+            cmd.extend(['--extra-vars', extra_vars])
+
+        cmd = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=False)
+        out, err = cmd.communicate()
+        rc = cmd.returncode
+        if rc:
+            raise errors.AnsibleError("%s: %s" % self.inventory_file, err)
+
+        try:
+            groups = utils.json_loads(out)
+        except:
+            raise errors.AnsibleError("invalid JSON response from script: %s" % self.inventory_file)
+
+        for (groupname, hostlist) in groups.iteritems():
+            for host in hostlist:
+                if host not in results:
+                    results.append(host)
+        return (results, groups)
+
+    # *****************************************************
+
+    def _get_variables_from_script(self, host, extra_vars=None):
+        ''' support per system variabes from external variable scripts, see web docs '''
+
+        cmd = [self.inventory_file, '--host', host]
+
+        if extra_vars:
+            cmd.extend(['--extra-vars', extra_vars])
+
+        cmd = subprocess.Popen(cmd,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            shell=False
+        )
+        out, err = cmd.communicate()
+
+        variables = {}
+        try:
+            variables = utils.json_loads(out)
+        except:
+            raise errors.AnsibleError("%s returned invalid result when called with hostname %s" % (
+                self.inventory_file,
+                host
+            ))
+        return variables
+
+    def _matches(self, host_name, pattern):
+        ''' returns if a hostname is matched by the pattern '''
+
+        # a pattern is in fnmatch format but more than one pattern
+        # can be strung together with semicolons. ex:
+        #   atlanta-web*.example.com;dc-web*.example.com
+
+        if host_name == '':
+            return False
+        pattern = pattern.replace(";",":")
+        subpatterns = pattern.split(":")
+        for subpattern in subpatterns:
+            if subpattern == 'all':
+                return True
+            if fnmatch.fnmatch(host_name, subpattern):
+                return True
+            elif subpattern in self.groups:
+                if host_name in self.groups[subpattern]:
+                    return True
+        return False

lib/ansible/playbook.py

@@ -17,6 +17,7 @@
 
 #############################################
 
+import ansible.inventory
 import ansible.runner
 import ansible.constants as C
 from ansible import utils
@@ -68,7 +69,6 @@ class PlayBook(object):
         if playbook is None or callbacks is None or runner_callbacks is None or stats is None:
             raise Exception('missing required arguments')
 
-        self.host_list        = host_list
         self.module_path      = module_path
         self.forks            = forks
         self.timeout          = timeout
@@ -88,9 +88,13 @@ class PlayBook(object):
         self.basedir          = os.path.dirname(playbook)
         self.playbook         = self._parse_playbook(playbook)
 
-        self.host_list, self.groups = ansible.runner.Runner.parse_hosts(
+        if override_hosts is not None:
-            host_list, override_hosts=self.override_hosts, extra_vars=self.extra_vars)
+            if type(override_hosts) != list:
-   
+                raise errors.AnsibleError("override hosts must be a list")
+            self.inventory = ansible.inventory.Inventory(override_hosts)
+        else:
+            self.inventory = ansible.inventory.Inventory(host_list)
+
     # *****************************************************
 
     def _get_vars(self, play, dirname):
@@ -233,7 +237,6 @@ class PlayBook(object):
     def _async_poll(self, runner, hosts, async_seconds, async_poll_interval, only_if):
         ''' launch an async job, if poll_interval is set, wait for completion '''
 
-        runner.host_list = hosts
         runner.background = async_seconds
         results = runner.run()
         self.stats.compute(results, poll=True)
@@ -257,7 +260,7 @@ class PlayBook(object):
             return results
 
         clock = async_seconds
-        runner.host_list = self.hosts_to_poll(results)
+        host_list = self.hosts_to_poll(results)
 
         poll_results = results
         while (clock >= 0):
@@ -267,11 +270,13 @@ class PlayBook(object):
             runner.module_name = 'async_status'
             runner.background  = 0  
             runner.pattern     = '*'
+            self.inventory.restrict_to(host_list)
             poll_results       = runner.run()
             self.stats.compute(poll_results, poll=True)
-            runner.host_list       = self.hosts_to_poll(poll_results)
+            host_list          = self.hosts_to_poll(poll_results)
+            self.inventory.lift_restriction()
 
-            if len(runner.host_list) == 0:
+            if len(host_list) == 0:
                 break
             if poll_results is None:
                 break
@@ -298,15 +303,16 @@ class PlayBook(object):
 
     # *****************************************************
 
-    def _run_module(self, pattern, host_list, module, args, vars, remote_user, 
+    def _run_module(self, pattern, module, args, vars, remote_user, 
         async_seconds, async_poll_interval, only_if, sudo, transport):
         ''' run a particular module step in a playbook '''
 
-        hosts = [ h for h in host_list if (h not in self.stats.failures) and (h not in self.stats.dark)]
+        hosts = [ h for h in self.inventory.list_hosts() if (h not in self.stats.failures) and (h not in self.stats.dark)]
+        self.inventory.restrict_to(hosts)
 
         runner = ansible.runner.Runner(
-            pattern=pattern, groups=self.groups, module_name=module,
+            pattern=pattern, inventory=self.inventory, module_name=module,
-            module_args=args, host_list=hosts, forks=self.forks,
+            module_args=args, forks=self.forks,
             remote_pass=self.remote_pass, module_path=self.module_path,
             timeout=self.timeout, remote_user=remote_user, 
             remote_port=self.remote_port, module_vars=vars,
@@ -317,13 +323,16 @@ class PlayBook(object):
         )
 
         if async_seconds == 0:
-            return runner.run()
+            results = runner.run()
         else:
-            return self._async_poll(runner, hosts, async_seconds, async_poll_interval, only_if)
+            results = self._async_poll(runner, hosts, async_seconds, async_poll_interval, only_if)
+
+        self.inventory.lift_restriction()
+        return results
 
     # *****************************************************
 
-    def _run_task(self, pattern=None, host_list=None, task=None, 
+    def _run_task(self, pattern=None, task=None, 
         remote_user=None, handlers=None, conditional=False, sudo=False, transport=None):
         ''' run a single task in the playbook and recursively run any subtasks.  '''
 
@@ -354,7 +363,7 @@ class PlayBook(object):
 
         # load up an appropriate ansible runner to
         # run the task in parallel
-        results = self._run_module(pattern, host_list, module_name, 
+        results = self._run_module(pattern, module_name, 
             module_args, module_vars, remote_user, async_seconds, 
             async_poll_interval, only_if, sudo, transport)
 
@@ -406,7 +415,7 @@ class PlayBook(object):
 
     # *****************************************************
 
-    def _do_conditional_imports(self, vars_files, host_list):
+    def _do_conditional_imports(self, vars_files):
         ''' handle the vars_files section, which can contain variables '''
         
         # FIXME: save parsed variable results in memory to avoid excessive re-reading/parsing
@@ -417,7 +426,7 @@ class PlayBook(object):
  
         if type(vars_files) != list:
             raise errors.AnsibleError("vars_files must be a list")
-        for host in host_list:
+        for host in self.inventory.list_hosts():
             cache_vars = SETUP_CACHE.get(host,{}) 
             SETUP_CACHE[host] = cache_vars
             for filename in vars_files:
@@ -460,16 +469,18 @@ class PlayBook(object):
 
         if vars_files is not None:
             self.callbacks.on_setup_secondary()
-            self._do_conditional_imports(vars_files, self.host_list)
+            self._do_conditional_imports(vars_files)
         else:
             self.callbacks.on_setup_primary()
 
-        host_list = [ h for h in self.host_list if not (h in self.stats.failures or h in self.stats.dark) ]
+        host_list = [ h for h in self.inventory.list_hosts(pattern) 
+                        if not (h in self.stats.failures or h in self.stats.dark) ]
+        self.inventory.restrict_to(host_list)
 
         # push any variables down to the system
         setup_results = ansible.runner.Runner(
-            pattern=pattern, groups=self.groups, module_name='setup',
+            pattern=pattern, module_name='setup',
-            module_args=vars, host_list=host_list,
+            module_args=vars, inventory=self.inventory,
             forks=self.forks, module_path=self.module_path,
             timeout=self.timeout, remote_user=user,
             remote_pass=self.remote_pass, remote_port=self.remote_port,
@@ -479,6 +490,8 @@ class PlayBook(object):
         ).run()
         self.stats.compute(setup_results, setup=True)
 
+        self.inventory.lift_restriction()
+
         # now for each result, load into the setup cache so we can
         # let runner template out future commands
         setup_ok = setup_results.get('contacted', {})
@@ -494,7 +507,6 @@ class PlayBook(object):
                     SETUP_CACHE[h].update(extra_vars)
                 except:
                     SETUP_CACHE[h] = extra_vars
-        return host_list
 
     # *****************************************************
 
@@ -530,7 +542,6 @@ class PlayBook(object):
         for task in tasks:
             self._run_task(
                 pattern=pattern,
-                host_list=self.host_list,
                 task=task, 
                 handlers=handlers,
                 remote_user=user,
@@ -547,16 +558,17 @@ class PlayBook(object):
         for task in handlers:
             triggered_by = task.get('run', None)
             if type(triggered_by) == list:
+                self.inventory.restrict_to(triggered_by)
                 self._run_task(
                    pattern=pattern, 
                    task=task,
                    handlers=[],
-                   host_list=triggered_by,
                    conditional=True,
                    remote_user=user,
                    sudo=sudo,
                    transport=transport
                 )
+                self.inventory.lift_restriction()
 
         # end of execution for this particular pattern.  Multiple patterns
         # can be in a single playbook file

lib/ansible/runner.py

@@ -18,7 +18,6 @@
 
 ################################################
 
-import fnmatch
 import multiprocessing
 import signal
 import os
@@ -27,10 +26,10 @@ import Queue
 import random
 import traceback
 import tempfile
-import subprocess
 
 import ansible.constants as C 
 import ansible.connection
+import ansible.inventory
 from ansible import utils
 from ansible import errors
 from ansible import callbacks as ans_callbacks
@@ -67,8 +66,6 @@ def _executor_hook(job_queue, result_queue):
 
 class Runner(object):
 
-    _external_variable_script = None
-
     def __init__(self, host_list=C.DEFAULT_HOST_LIST, module_path=C.DEFAULT_MODULE_PATH,
         module_name=C.DEFAULT_MODULE_NAME, module_args=C.DEFAULT_MODULE_ARGS, 
         forks=C.DEFAULT_FORKS, timeout=C.DEFAULT_TIMEOUT, pattern=C.DEFAULT_PATTERN,
@@ -76,7 +73,7 @@ class Runner(object):
         sudo_pass=C.DEFAULT_SUDO_PASS, remote_port=C.DEFAULT_REMOTE_PORT, background=0, 
         basedir=None, setup_cache=None, transport=C.DEFAULT_TRANSPORT, 
         conditional='True', groups={}, callbacks=None, verbose=False,
-        debug=False, sudo=False, extra_vars=None, module_vars=None):
+        debug=False, sudo=False, extra_vars=None, module_vars=None, inventory=None):
    
         if setup_cache is None:
             setup_cache = {}
@@ -92,11 +89,10 @@ class Runner(object):
         self.transport = transport
         self.connector = ansible.connection.Connection(self, self.transport)
 
-        if type(host_list) == str:
+        if inventory is None:
-            self.host_list, self.groups = self.parse_hosts(host_list)
+            self.inventory = ansible.inventory.Inventory(host_list, extra_vars)
         else:
-            self.host_list = host_list
+            self.inventory = inventory
-            self.groups    = groups
 
         self.setup_cache = setup_cache
         self.conditional = conditional
@@ -130,106 +126,6 @@ class Runner(object):
 
     # *****************************************************
 
-    @classmethod
-    def parse_hosts_from_regular_file(cls, host_list):
-        ''' parse a textual host file '''
-
-        results = []
-        groups = dict(ungrouped=[])
-        lines = file(host_list).read().split("\n")
-        group_name = 'ungrouped'
-        for item in lines:
-            item = item.lstrip().rstrip()
-            if item.startswith("#"):
-                # ignore commented out lines
-                pass
-            elif item.startswith("["):
-                # looks like a group
-                group_name = item.replace("[","").replace("]","").lstrip().rstrip()
-                groups[group_name] = []
-            elif item != "":
-                # looks like a regular host
-                groups[group_name].append(item)
-                if not item in results:
-                    results.append(item)
-        return (results, groups)
-
-    # *****************************************************
-
-    @classmethod
-    def parse_hosts_from_script(cls, host_list, extra_vars):
-        ''' evaluate a script that returns list of hosts by groups '''
-
-        results = []
-        groups = dict(ungrouped=[])
-        host_list = os.path.abspath(host_list)
-        cls._external_variable_script = host_list
-        cmd = [host_list, '--list']
-        if extra_vars:
-            cmd.extend(['--extra-vars', extra_vars])
-        cmd = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=False)
-        out, err = cmd.communicate()
-        rc = cmd.returncode
-        if rc:
-            raise errors.AnsibleError("%s: %s" % (host_list, err))
-        try:
-            groups = utils.json_loads(out)
-        except:
-            raise errors.AnsibleError("invalid JSON response from script: %s" % host_list)
-        for (groupname, hostlist) in groups.iteritems():
-            for host in hostlist:
-                if host not in results:
-                    results.append(host)
-        return (results, groups)
-
-    # *****************************************************
-
-    @classmethod
-    def parse_hosts(cls, host_list, override_hosts=None, extra_vars=None):
-        ''' parse the host inventory file, returns (hosts, groups) '''
-
-        if override_hosts is not None:
-            if type(override_hosts) != list:
-                raise errors.AnsibleError("override hosts must be a list")
-            return (override_hosts, dict(ungrouped=override_hosts))
-
-        if type(host_list) == list:
-            raise Exception("function can only be called on inventory files")
-
-        host_list = os.path.expanduser(host_list)
-        if not os.path.exists(host_list):
-            raise errors.AnsibleFileNotFound("inventory file not found: %s" % host_list)
-
-        if not os.access(host_list, os.X_OK):
-            return Runner.parse_hosts_from_regular_file(host_list)
-        else:
-            return Runner.parse_hosts_from_script(host_list, extra_vars)
-
-    # *****************************************************
-
-    def _matches(self, host_name, pattern):
-        ''' returns if a hostname is matched by the pattern '''
-
-        # a pattern is in fnmatch format but more than one pattern
-        # can be strung together with semicolons. ex:
-        #   atlanta-web*.example.com;dc-web*.example.com
-
-        if host_name == '':
-            return False
-        pattern = pattern.replace(";",":")
-        subpatterns = pattern.split(":")
-        for subpattern in subpatterns:
-            if subpattern == 'all':
-                return True
-            if fnmatch.fnmatch(host_name, subpattern):
-                return True
-            elif subpattern in self.groups:
-                if host_name in self.groups[subpattern]:
-                    return True
-        return False
-
-    # *****************************************************
-
     def _connect(self, host):
         ''' connects to a host, returns (is_successful, connection_object OR traceback_string) '''
 
@@ -296,34 +192,6 @@ class Runner(object):
 
     # *****************************************************
 
-    def _add_variables_from_script(self, conn, inject):
-        ''' support per system variabes from external variable scripts, see web docs '''
-
-        host = conn.host
-
-        cmd = [Runner._external_variable_script, '--host', host]
-        if self.extra_vars:
-            cmd.extend(['--extra-vars', self.extra_vars])
-
-        cmd = subprocess.Popen(cmd,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.PIPE,
-            shell=False
-        )
-        out, err = cmd.communicate()
-        inject2 = {}
-        try:
-            inject2 = utils.json_loads(out)
-        except:
-            raise errors.AnsibleError("%s returned invalid result when called with hostname %s" % (
-                Runner._external_variable_script,
-                host
-            ))
-        # store injected variables in the templates
-        inject.update(inject2)
-
-    # *****************************************************
-
     def _add_setup_vars(self, inject, args):
         ''' setup module variables need special handling '''
 
@@ -377,8 +245,9 @@ class Runner(object):
         if not eval(conditional):
             return [ utils.smjson(dict(skipped=True)), None, 'skipped' ]
 
-        if Runner._external_variable_script is not None:
+        host_variables = self.inventory.get_variables(conn.host, self.extra_vars)
-            self._add_variables_from_script(conn, inject)
+        inject.update(host_variables)
+
         if self.module_name == 'setup':
             args = self._add_setup_vars(inject, args)
             args = self._add_setup_metadata(args)
@@ -692,13 +561,6 @@ class Runner(object):
 
     # *****************************************************
 
-    def _match_hosts(self, pattern):
-        ''' return all matched hosts fitting a pattern '''
-
-        return [ h for h in self.host_list if self._matches(h, pattern) ]
-
-    # *****************************************************
-
     def _parallel_exec(self, hosts):
         ''' handles mulitprocessing when more than 1 fork is required '''
 
@@ -745,7 +607,7 @@ class Runner(object):
                 results2["dark"][host] = result
 
         # hosts which were contacted but never got a chance to return
-        for host in self._match_hosts(self.pattern):
+        for host in self.inventory.list_hosts(self.pattern):
             if not (host in results2['dark'] or host in results2['contacted']):
                 results2["dark"][host] = {}
 
@@ -757,7 +619,7 @@ class Runner(object):
         ''' xfer & run module on all matched hosts '''
        
         # find hosts that match the pattern
-        hosts = self._match_hosts(self.pattern)
+        hosts = self.inventory.list_hosts(self.pattern)
         if len(hosts) == 0:
             self.callbacks.on_no_hosts()
             return dict(contacted={}, dark={})