Add top-level python directory to opensource repo

This adds a home in our opensource repo for python libraries and binaries,
under a top-level "python" directory.  Future CLs will relocate ICANN
reporting bits and pieces to new homes under this directory, and will use
the MOE configuration and python_directory_import rule defined here.

This approach is roughly modeled on the protobuf Bazel opensource project,
which also uses a top-level directory for various languages, and also uses
the "imports" parameter to exclude that directory in python module names:
  https://github.com/google/protobuf/blob/v3.0.0-beta-3/BUILD#L568
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=127459882
This commit is contained in:
nickfelt 2016-07-14 12:22:51 -07:00 committed by Ben McIlwain
parent 4ccc016e5c
commit eed319990a
16 changed files with 359 additions and 43 deletions

View file

@ -14,17 +14,22 @@
package google.registry.xml;
import static com.google.common.base.MoreObjects.firstNonNull;
import static com.google.common.base.Preconditions.checkNotNull;
import static com.google.common.truth.Truth.assert_;
import static google.registry.util.DiffUtils.prettyPrintDeepDiff;
import static org.joda.time.DateTimeZone.UTC;
import com.google.common.base.Splitter;
import com.google.common.collect.ImmutableList;
import com.google.common.collect.ImmutableMap;
import com.google.common.collect.ImmutableSet;
import com.google.common.collect.Iterables;
import com.google.common.net.InetAddresses;
import com.google.common.net.InternetDomainName;
import java.util.AbstractMap;
import java.util.HashMap;
import java.util.HashSet;
import java.util.List;
import java.util.Map;
import java.util.Set;
import javax.annotation.Nullable;
@ -58,91 +63,205 @@ public class XmlTestUtils {
}
}
/** Deeply explore the object and normalize values so that things we consider equal compare so. */
private static Object normalize(Object obj, @Nullable String path, Set<String> ignoredPaths)
throws Exception {
/**
* Map an element or attribute name using a namespace map to replace the namespace identifier
* with the complete URI as given in the map. If the name has no namespace identifier, the default
* namespace mapping is used. If the namespace identifier does not exist in the map, the name is
* left unchanged.
*/
private static String mapName(
@Nullable String name, Map<String, String> nsMap, boolean mapDefaultNamespace) {
if (name == null) {
return null;
}
String ns;
String simpleKey;
List<String> components = Splitter.on(':').splitToList(name);
// Handle names without identifiers, meaning they are in the default namespace.
if (components.size() < 2) {
if (!mapDefaultNamespace) {
return name;
}
ns = "";
simpleKey = name;
// Handle names with identifiers.
} else {
ns = components.get(0);
simpleKey = components.get(1);
}
// If the map does not contain the specified identifier (or "" for the default), don't do
// anything.
if (nsMap.containsKey(ns)) {
ns = nsMap.get(ns);
}
return ns.isEmpty() ? simpleKey : (ns + ':' + simpleKey);
}
/**
* Deeply explore the object and normalize values so that things we consider equal compare so.
* The return value consists of two parts: the updated key and the value. The value is
* straightforward enough: it is the rendering of the subtree to be attached at the current point.
* The key is more complicated, because of namespaces. When an XML element specifies namespaces
* using xmlns attributes, those namespaces apply to the element as well as all of its
* descendants. That means that, when prefixing the element name with the full namespace path,
* as required to do proper comparison, the element name depends on its children. When looping
* through a JSONObject map, we can't just recursively generate the value and store it using the
* key. We may have to update the key as well, to get the namespaces correct. A returned key of
* null indicates that we should use the existing key. A non-null key indicates that we should
* replace the existing key.
*
* @param elementName the name under which the current subtree was found, or null if the current
* subtree's name is nonexistent or irrelevant
* @param obj the current subtree
* @param path the (non-namespaced) element path used for ignoredPaths purposes
* @param ignoredPaths the set of paths whose values should be set to IGNORED
* @param nsMap the inherited namespace identifier-to-URI map
* @return the key under which the rendered subtree should be stored (or null), and the rendered
* subtree
*/
private static Map.Entry<String, Object> normalize(
@Nullable String elementName,
Object obj,
@Nullable String path,
Set<String> ignoredPaths,
Map<String, String> nsMap) throws Exception {
if (obj instanceof JSONObject) {
JSONObject jsonObject = (JSONObject) obj;
Map<String, Object> map = new HashMap<>();
// getNames helpfully returns null rather than empty when there are no names.
for (String key : firstNonNull(JSONObject.getNames(jsonObject), new String[0])) {
// Recursively transform json maps, remove namespaces and the "xmlns" key.
if (key.startsWith("xmlns")) {
continue;
String[] names = JSONObject.getNames(jsonObject);
if (names != null) {
// Separate all elements and keys into namespace specifications, which we must process
// first, and everything else.
ImmutableList.Builder<String> namespacesBuilder = new ImmutableList.Builder<>();
ImmutableList.Builder<String> othersBuilder = new ImmutableList.Builder<>();
for (String key : names) {
(key.startsWith("xmlns") ? namespacesBuilder : othersBuilder).add(key);
}
String simpleKey = key.replaceAll(".*:", "");
String newPath = path == null ? simpleKey : path + "." + simpleKey;
Object value;
if (ignoredPaths.contains(newPath)) {
// Set ignored fields to a value that will compare equal.
value = "IGNORED";
} else {
value = normalize(jsonObject.get(key), newPath, ignoredPaths);
// First, handle all namespace specifications, updating our ns-to-URI map. Use a HashMap
// rather than an ImmutableMap.Builder so that we can override existing map entries.
HashMap<String, String> newNsMap = new HashMap<>();
newNsMap.putAll(nsMap);
for (String key : namespacesBuilder.build()) {
// Parse the attribute name, of the form xmlns:nsid, and extract the namespace identifier.
// If there's no colon, we are setting the default namespace.
List<String> components = Splitter.on(':').splitToList(key);
String ns = (components.size() >= 2) ? components.get(1) : "";
newNsMap.put(ns, jsonObject.get(key).toString());
}
nsMap = ImmutableMap.copyOf(newNsMap);
// Now, handle the non-namespace items, recursively transforming the map and mapping all
// namespaces to the full URI for proper comparison.
for (String key : othersBuilder.build()) {
String simpleKey = Iterables.getLast(Splitter.on(':').split(key));
String newPath = (path == null) ? simpleKey : (path + "." + simpleKey);
String mappedKey;
Object value;
if (ignoredPaths.contains(newPath)) {
mappedKey = null;
// Set ignored fields to a value that will compare equal.
value = "IGNORED";
} else {
Map.Entry<String, Object> simpleEntry =
normalize(key, jsonObject.get(key), newPath, ignoredPaths, nsMap);
mappedKey = simpleEntry.getKey();
value = simpleEntry.getValue();
}
if (mappedKey == null) {
// Note that this does not follow the XML rules exactly. I read somewhere that attribute
// names, unlike element names, never use the default namespace. But after
// JSONification, we cannot distinguish between attributes and child elements, so we
// apply the default namespace to everything. Hopefully that will not cause a problem.
mappedKey = key.equals("content") ? key : mapName(key, nsMap, true);
}
map.put(mappedKey, value);
}
map.put(simpleKey, value);
}
// Map the namespace of the element name of the map we are normalizing.
elementName = mapName(elementName, nsMap, true);
// If a node has both text content and attributes, the text content will end up under a key
// called "content". If that's the only thing left (which will only happen if there was an
// "xmlns:*" key that we removed), treat the node as just text and recurse.
if (map.size() == 1 && map.get("content") != null) {
return normalize(jsonObject.get("content"), path, ignoredPaths);
if (map.size() == 1 && map.containsKey("content")) {
return new AbstractMap.SimpleEntry<>(
elementName,
normalize(null, jsonObject.get("content"), path, ignoredPaths, nsMap).getValue());
}
// The conversion to JSON converts <a/> into "" and the semantically equivalent <a></a> into
// an empty map, so normalize that here.
return map.isEmpty() ? "" : map;
return new AbstractMap.SimpleEntry<>(elementName, map.isEmpty() ? "" : map);
}
if (obj instanceof JSONArray) {
// Another problem resulting from JSONification: If the array contains elements whose names
// are the same before URI expansion, but different after URI expansion, because they use
// xmlns attribute that define the namespaces differently, we will screw up. Again, hopefully
// that doesn't happen much. The reverse is also true: If the array contains names that are
// different before URI expansion, but the same after, we may have a problem, because the
// elements will wind up in different JSONArrays as a result of JSONification. We wave our
// hands and just assume that the URI expansion of the first element holds for all others.
Set<Object> set = new HashSet<>();
String mappedKey = null;
for (int i = 0; i < ((JSONArray) obj).length(); ++i) {
set.add(normalize(((JSONArray) obj).get(i), path, ignoredPaths));
Map.Entry<String, Object> simpleEntry =
normalize(null, ((JSONArray) obj).get(i), path, ignoredPaths, nsMap);
if (i == 0) {
mappedKey = simpleEntry.getKey();
}
set.add(simpleEntry.getValue());
}
return set;
return new AbstractMap.SimpleEntry<String, Object>(mappedKey, set);
}
if (obj instanceof Number) {
return obj.toString();
return new AbstractMap.SimpleEntry<String, Object>(null, obj.toString());
}
if (obj instanceof Boolean) {
return ((Boolean) obj) ? "1" : "0";
return new AbstractMap.SimpleEntry<String, Object>(null, ((Boolean) obj) ? "1" : "0");
}
if (obj instanceof String) {
// Turn stringified booleans into integers. Both are acceptable as xml boolean values, but
// we use "true" and "false" whereas the samples use "1" and "0".
if (obj.equals("true")) {
return "1";
return new AbstractMap.SimpleEntry<String, Object>(null, "1");
}
if (obj.equals("false")) {
return "0";
return new AbstractMap.SimpleEntry<String, Object>(null, "0");
}
String string = obj.toString();
// We use a slightly different datetime format (both legal) than the samples, so normalize
// both into Datetime objects.
try {
return ISODateTimeFormat.dateTime().parseDateTime(string).toDateTime(UTC);
return new AbstractMap.SimpleEntry<String, Object>(
null, ISODateTimeFormat.dateTime().parseDateTime(string).toDateTime(UTC));
} catch (IllegalArgumentException e) {
// It wasn't a DateTime.
}
try {
return ISODateTimeFormat.dateTimeNoMillis().parseDateTime(string).toDateTime(UTC);
return new AbstractMap.SimpleEntry<String, Object>(
null, ISODateTimeFormat.dateTimeNoMillis().parseDateTime(string).toDateTime(UTC));
} catch (IllegalArgumentException e) {
// It wasn't a DateTime.
}
try {
if (!InternetDomainName.isValid(string)) {
// It's not a domain name, but it is an InetAddress. Ergo, it's an ip address.
return InetAddresses.forString(string);
return new AbstractMap.SimpleEntry<String, Object>(null, InetAddresses.forString(string));
}
} catch (IllegalArgumentException e) {
// Not an ip address.
}
return string;
return new AbstractMap.SimpleEntry<String, Object>(null, string);
}
return checkNotNull(obj);
return new AbstractMap.SimpleEntry<>(null, checkNotNull(obj));
}
@SuppressWarnings("unchecked")
private static Map<String, Object> toComparableJson(
String xml, String... ignoredPaths) throws Exception {
return (Map<String, Object>) normalize(
XML.toJSONObject(xml), null, ImmutableSet.copyOf(ignoredPaths));
null,
XML.toJSONObject(xml),
null,
ImmutableSet.copyOf(ignoredPaths),
ImmutableMap.<String, String>of()).getValue();
}
}