jjfiv

import ciir.jfoley.chai.time.Debouncer;
import org.apache.lucene.analysis.en.EnglishAnalyzer;
import org.apache.lucene.document.Field;
import org.apache.lucene.document.StringField;
import org.apache.lucene.document.TextField;
import org.apache.lucene.index.IndexWriter;
import org.apache.lucene.index.IndexWriterConfig;
import org.apache.lucene.store.FSDirectory;
import org.jsoup.Jsoup;

jjfiv

import socket
host = socket.gethostname()
onSydney=(host == 'sydney.cs.umass.edu')

if onSydney:
    print('CRFSUITE:=/mnt/nfs/work3/jfoley/bin/crfsuite-0.12/bin/crfsuite')
    print("PREFIX=qsub -b y -cwd -sync y -l mem_free=8G -l mem_token=8G -o $@.out -e $@.err ")
    print("JAVA:=/mnt/nfs/work3/jfoley/bin/jdk1.8.0_31/bin/java -ea -Xmx7G")
    print("SUFFIX:=") # log file created through qsub
else:

jjfiv

public static double computeAP(List<Boolean> isTruePositiveFromRanking, int numRelevant) {
    // if there are no relevant documents,
    // the average is artificially defined as zero, to mimic trec_eval
    // Really, the output is NaN, or the query should be ignored [point of debate]
    if(numRelevant == 0) return 0;

    double sumPrecision = 0;
    int recallPointCount = 0;

    for (int i = 0; i < data.size(); i++) {

jjfiv

def ternary(bool, pos, neg):
    if bool:
        return pos
    else:
        return neg

jjfiv

import numpy as np
import torch
from torch.nn.utils.rnn import pack_padded_sequence, pad_sequence

def pack_lstm(items, lstm):
    N = len(items)
    reorder_args = np.argsort([len(it) for it in items])[::-1]
    origin_args = torch.from_numpy(np.argsort(reorder_args))
    ordered = [items[i] for i in reorder_args]
    packed_items = pack_padded_sequence(pad_sequence(ordered, batch_first=True), [len(od) for od in ordered], batch_first=True)

jjfiv

import agent_ql as aq

# Diaz, F. "Condensed List Relevance Models." (ICTIR 2015)
def CLRM3(query, originalWeight=0.3):
    first_pass = aq.ql(aq.tokenize(query))
    RM = aq.term_probability_model()
    for doc in first_pass.search_now():
        RM += doc.to_term_probabilities() * doc.score
        return first_pass.results().re_rank( first_pass.mixture_model(RM, originalWeight) )

jjfiv

import java.util.Random;
import java.util.Scanner;

// We discussed academic honesty, so when you re-type this code, be sure to cite it in a comment!
public class GuessingGame {

	/**
	 * A Java program will run code in a special ``main`` method.
	 * Note that Java has two types of comments: block (slash-star ... star-slash), and line ("slash-slash") comments.
	 * For now we ignore args, which is an array of strings that the user might have passed in.

jjfiv

package edu.smith.cs.csc262.coopsh.apps;

import edu.smith.cs.csc262.coopsh.ShellEnvironment;
import edu.smith.cs.csc262.coopsh.Task;

/**
 * This is a full implementation of Echo.
 * @author jfoley
 *
 */

jjfiv

import java.io.*;
import java.util.ArrayList;
import java.util.List;
import java.util.stream.Collectors;
import java.util.zip.ZipEntry;
import java.util.zip.ZipFile;
import java.util.zip.ZipOutputStream;

public class ZipSplit {

jjfiv

use std::convert::TryInto;
use tantivy::{Searcher, Term};

#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct CountStats {
    pub collection_frequency: u64,
    pub document_frequency: u64,
    pub collection_length: u64,
    pub document_count: u64,
}