1 /**
2 * Licensed to the Apache Software Foundation (ASF) under one or more
3 * contributor license agreements. See the NOTICE file distributed with
4 * this work for additional information regarding copyright ownership.
5 * The ASF licenses this file to You under the Apache License, Version 2.0
6 * (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 *
9 * http://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17
18 package org.apache.oodt.pcs.pedigree;
19
20 //OODT imports
21 import org.apache.oodt.pcs.metadata.PCSMetadata;
22 import org.apache.oodt.pcs.metadata.PCSConfigMetadata;
23 import org.apache.oodt.pcs.query.FilenameQuery;
24 import org.apache.oodt.pcs.query.InputFilesQuery;
25 import org.apache.oodt.pcs.query.JobIdQuery;
26 import org.apache.oodt.pcs.util.FileManagerUtils;
27 import org.apache.oodt.cas.filemgr.structs.Product;
28 import org.apache.oodt.cas.metadata.Metadata;
29
30 //JDK imports
31 import java.util.Iterator;
32 import java.util.List;
33 import java.util.Stack;
34 import java.util.Vector;
35
36 /**
37 *
38 * A class to provide pedigre tracking for PCS {@link Product}s.
39 *
40 * @author mattmann
41 * @version $Revision$
42 */
43 public class Pedigree implements PCSMetadata, PCSConfigMetadata {
44
45 /* our file manager interface */
46 private FileManagerUtils fm;
47
48 /* should we include not cataloged products in the pedigree? */
49 private boolean listNotCataloged = false;
50
51 /* are there any product types that we should exclude from the pedigree? */
52 private List prodTypeExcludeList;
53
54 /**
55 *
56 * Constructs a new Pedigree object that will connect to the file manager at
57 * the given {@link URL} specified by the <code>fmUrlStr</code> string
58 * parameter.
59 *
60 * @param fmUrlStr
61 * The String {@link URL} of the file manager to connect to.
62 * @param listNotCat
63 * Whether or not we should include non-cataloged products in the
64 * pedigree.
65 * @param excludeList
66 * A {@link List} of String {@link ProductType} names that should be
67 * excluded from the Pedigree.
68 */
69 public Pedigree(String fmUrlStr, boolean listNotCat, List excludeList) {
70 this(new FileManagerUtils(FileManagerUtils.safeGetUrlFromString(fmUrlStr)),
71 listNotCat, excludeList);
72
73 }
74
75 /**
76 *
77 * Constructs a new Pedigree object that will connect to the file manager
78 * specified by the {@link FileManagerUtils} object passed in.
79 *
80 * @param fm
81 * The PCS interface to the File Manager.
82 * @param listNotCat
83 * Whether or not we should include non-cataloged products in the
84 * pedigree.
85 * @param excludeList
86 * A {@link List} of String {@link ProductType} names that should be
87 * excluded from the Pedigree.
88 */
89 public Pedigree(FileManagerUtils fm, boolean listNotCat, List excludeList) {
90 this.fm = fm;
91 this.listNotCataloged = listNotCat;
92 this.prodTypeExcludeList = excludeList;
93 }
94
95 /**
96 * Performs a full pedigree of the specified {@link Product} <code>orig</code>
97 * . If <code>upstream</code> is set to true, an upstream pedigree is
98 * performed, otherwise, a downstream pedigree is performed.
99 *
100 * @param orig
101 * The {@link Product} to perform a pedigree of.
102 * @param upstream
103 * Whether or not we should do an upstream (true) or downstream
104 * (false) pedigree.
105 * @return A {@link PedigreeTree} containing the Pedigree of a given product.
106 */
107 public PedigreeTree doPedigree(Product orig, boolean upstream) {
108 List pedProds = null;
109 PedigreeTreeNode origRoot = PedigreeTreeNode
110 .getPedigreeTreeNodeFromProduct(orig, null);
111
112 // System.out.println("Doing pedigree: upstream: ["+upstream+"]");
113 Stack roots = new Stack();
114 roots.add(origRoot);
115
116 do {
117
118 PedigreeTreeNode currRoot = (PedigreeTreeNode) roots.pop();
119 /*
120 * System.out.println("Examining root: ["+currRoot.getNodeProduct().
121 * getProductName()+"]");
122 */
123
124 if (upstream) {
125 pedProds = getUpstreamPedigreedProducts(currRoot.getNodeProduct());
126 } else {
127 pedProds = getDownstreamPedigreedProducts(currRoot.getNodeProduct());
128 }
129
130 if (pedProds != null && pedProds.size() > 0) {
131 for (Iterator i = pedProds.iterator(); i.hasNext();) {
132 Product p = (Product) i.next();
133 if (p.getProductName().equals(
134 currRoot.getNodeProduct().getProductName())) {
135 // don't allow for the same pedigreed product to be
136 // added to the list
137 continue;
138 }
139 PedigreeTreeNode prodNode = PedigreeTreeNode
140 .getPedigreeTreeNodeFromProduct(p, currRoot);
141 roots.add(prodNode);
142 }
143 } else {
144 // System.out.println("Pedigree products is null or size 0");
145 }
146
147 } while (!roots.empty());
148
149 return new PedigreeTree(origRoot);
150 }
151
152 /**
153 * Returns the most direct ancestors (a {@link List} of {@link Product}s)
154 * upstream from the given {@link Product} named <code>orig</code>.
155 *
156 * @param orig
157 * The {@link Product} to get direct upstream relatives of.
158 * @return A {@link List} of {@link Product}s directly upstream from the given
159 * {@link Product}.
160 */
161 public List getUpstreamPedigreedProducts(Product orig) {
162 if (orig == null
163 || (orig != null && orig.getProductType() == null)
164 || (orig != null && orig.getProductType() != null
165 && orig.getProductType().getName() == null)
166 || (orig != null && orig.getProductType() != null
167 && orig.getProductType().getName() != null &&
168 orig.getProductType().getName().equals(UNKNOWN))) {
169 return new Vector();
170 }
171 Metadata pMet = fm.safeGetMetadata(orig);
172 return getProducts(pMet.getAllMetadata(INPUT_FILES));
173
174 }
175
176 /**
177 * Returns the most direct ancestors (a {@link List} of {@link Product}s)
178 * downstream from the given {@link Product} named <code>orig</code>.
179 *
180 * @param orig
181 * The {@link Product} to get direct downstream relatives of.
182 * @return A {@link List} of {@link Product}s directly downstream from the
183 * given {@link Product}.
184 */
185 public List getDownstreamPedigreedProducts(Product orig) {
186 return fm.queryAllTypes(new InputFilesQuery(orig.getProductName(), fm)
187 .buildQuery(), this.prodTypeExcludeList);
188 }
189
190 /**
191 * Gets all associated {@link Product}s with the provided
192 * {@link WorkflowInstance} identified by its ID (the <code>wInstId</code>
193 * string parameter).
194 *
195 * @param wInstId
196 * The ID of the {@link WorkflowInstance} to look up the
197 * {@link Product}s for.
198 * @return A {@link List} of {@link Product}s associated with the provided
199 * Workflow Instance ID.
200 */
201 public List getWorkflowInstProds(String wInstId) {
202 return fm.queryAllTypes(new JobIdQuery(wInstId, fm).buildQuery(),
203 this.prodTypeExcludeList);
204 }
205
206 private List getProducts(List prodNames) {
207 if (prodNames == null || (prodNames != null && prodNames.size() == 0)) {
208 return new Vector();
209 }
210
211 List prods = new Vector(prodNames.size());
212
213 for (Iterator i = prodNames.iterator(); i.hasNext();) {
214 String prodName = (String) i.next();
215 List prodList = fm.queryAllTypes(new FilenameQuery(prodName, fm)
216 .buildQuery(), this.prodTypeExcludeList);
217 if (prodList != null && prodList.size() > 0) {
218 prods.add((Product) prodList.get(0));
219 } else {
220 if (this.listNotCataloged) {
221 // create a new product and add it
222 prods.add(Product.getDefaultFlatProduct(prodName, UNKNOWN));
223 }
224 }
225 }
226
227 return prods;
228 }
229
230 }