hivemall-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From takuti <...@git.apache.org>
Subject [GitHub] incubator-hivemall pull request #97: [HIVEMALL-130] Support user-defined dic...
Date Wed, 12 Jul 2017 04:09:28 GMT
Github user takuti commented on a diff in the pull request:

    https://github.com/apache/incubator-hivemall/pull/97#discussion_r126860711
  
    --- Diff: nlp/src/main/java/hivemall/nlp/tokenizer/KuromojiUDF.java ---
    @@ -170,6 +192,100 @@ private static CharArraySet stopWords(@Nonnull final String[] array)
             return results;
         }
     
    +    @Nullable
    +    private static UserDictionary userDictionary(@Nonnull final ObjectInspector oi)
    +            throws UDFArgumentException {
    +        if (HiveUtils.isConstListOI(oi)) {
    +            return userDictionary(HiveUtils.getConstStringArray(oi));
    +        } else if (HiveUtils.isConstString(oi)) {
    +            return userDictionary(HiveUtils.getConstString(oi));
    +        } else {
    +            throw new UDFArgumentException(
    +                "User dictionary MUST be given as an array of constant string or constant
string (URL)");
    +        }
    +    }
    +
    +    @Nullable
    +    private static UserDictionary userDictionary(@Nullable final String[] userDictArray)
    +            throws UDFArgumentException {
    +        if (userDictArray == null) {
    +            return null;
    +        }
    +
    +        StringBuilder builder = new StringBuilder();
    +        for (String row : userDictArray) {
    +            builder.append(row);
    +            builder.append("\r\n");
    +        }
    +        InputStream is = new ByteArrayInputStream(builder.toString().getBytes());
    +        final Reader reader = new InputStreamReader(is);
    +        try {
    +            return UserDictionary.open(reader); // return null if empty
    +        } catch (Throwable e) {
    +            throw new UDFArgumentException(
    +                "Failed to create user dictionary based on the given array<string>:
" + e);
    +        }
    +    }
    +
    +    @Nullable
    +    private static UserDictionary userDictionary(@Nullable final String userDictURL)
    +            throws UDFArgumentException {
    +        if (userDictURL == null) {
    +            return null;
    +        }
    +
    +        final HttpURLConnection conn;
    +        try {
    +            conn = HttpUtils.getHttpURLConnection(userDictURL);
    +        } catch (Throwable e) {
    +            throw new UDFArgumentException("Failed to create HTTP connection to the URL:
" + e);
    +        }
    +
    +        // allow to read as a compressed GZIP file for efficiency
    +        conn.setRequestProperty("Accept-Encoding", "gzip");
    +
    +        conn.setConnectTimeout(CONNECT_TIMEOUT_MS); // throw exception from connect()
    +        conn.setReadTimeout(READ_TIMEOUT_MS); // throw exception from getXXX() methods
    +
    +        final int responseCode;
    +        try {
    +            responseCode = conn.getResponseCode();
    +        } catch (Throwable e) {
    +            throw new UDFArgumentException("Failed to get response code: " + e);
    +        }
    +        if (responseCode != 200) {
    +            throw new UDFArgumentException("God invalid response code: " + responseCode);
    +        }
    +
    +        final String contentType = conn.getContentType();
    +        boolean textContent = contentType.startsWith("text/plain");
    +        boolean binaryContent = contentType.startsWith("application/octet-stream");
    +        if (!textContent && !binaryContent) {
    +            throw new UDFArgumentException(
    +                "User dictionary URL indicates unexpected content type: " + contentType);
    +        }
    +
    +        InputStream is;
    +        try {
    +            is = HttpUtils.getBoundedInputStream(conn, MAX_INPUT_STREAM_SIZE);
    +            if ((textContent && "gzip".equals(conn.getContentEncoding())) ||
binaryContent) {
    +                is = new GZIPInputStream(is);
    --- End diff --
    
    Since URL can be easily impersonated, I don't accept to use url string for the purpose
of detecting file type. However, for double-check purpose, your suggestion `if(urlString.endWith(".gz"))`
is partially useful.


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message