org.apache.nutch.crawl
Interface FetchSchedule

All Superinterfaces:

org.apache.hadoop.conf.Configurable

All Known Implementing Classes:

AbstractFetchSchedule, AdaptiveFetchSchedule, DefaultFetchSchedule

public interface FetchSchedule
extends org.apache.hadoop.conf.Configurable

This interface defines the contract for implementations that manipulate fetch times and re-fetch intervals.

Author:

Andrzej Bialecki

Field Summary

static int	SECONDS_PER_DAY
static int	STATUS_MODIFIED Page is known to have been modified since our last visit.
static int	STATUS_NOTMODIFIED Page is known to remain unmodified since our last visit.
static int	STATUS_UNKNOWN It is unknown whether page was changed since our last visit.

Method Summary

long	calculateLastFetchTime(CrawlDatum datum) Calculates last fetch time of the given CrawlDatum.
CrawlDatum	forceRefetch(org.apache.hadoop.io.Text url, CrawlDatum datum, boolean asap) This method resets fetchTime, fetchInterval, modifiedTime and page signature, so that it forces refetching.
CrawlDatum	initializeSchedule(org.apache.hadoop.io.Text url, CrawlDatum datum) Initialize fetch schedule related data.
CrawlDatum	setFetchSchedule(org.apache.hadoop.io.Text url, CrawlDatum datum, long prevFetchTime, long prevModifiedTime, long fetchTime, long modifiedTime, int state) Sets the fetchInterval and fetchTime on a successfully fetched page.
CrawlDatum	setPageGoneSchedule(org.apache.hadoop.io.Text url, CrawlDatum datum, long prevFetchTime, long prevModifiedTime, long fetchTime) This method specifies how to schedule refetching of pages marked as GONE.
CrawlDatum	setPageRetrySchedule(org.apache.hadoop.io.Text url, CrawlDatum datum, long prevFetchTime, long prevModifiedTime, long fetchTime) This method adjusts the fetch schedule if fetching needs to be re-tried due to transient errors.
boolean	shouldFetch(org.apache.hadoop.io.Text url, CrawlDatum datum, long curTime) This method provides information whether the page is suitable for selection in the current fetchlist.

Methods inherited from interface org.apache.hadoop.conf.Configurable

getConf, setConf

Field Detail

STATUS_UNKNOWN

static final int STATUS_UNKNOWN

It is unknown whether page was changed since our last visit.

See Also:

Constant Field Values

STATUS_MODIFIED

static final int STATUS_MODIFIED

Page is known to have been modified since our last visit.

See Also:

Constant Field Values

STATUS_NOTMODIFIED

static final int STATUS_NOTMODIFIED

Page is known to remain unmodified since our last visit.

See Also:

Constant Field Values

SECONDS_PER_DAY

static final int SECONDS_PER_DAY

See Also:

Constant Field Values

Method Detail

initializeSchedule

CrawlDatum initializeSchedule(org.apache.hadoop.io.Text url,
                              CrawlDatum datum)

Initialize fetch schedule related data. Implementations should at least set the fetchTime and fetchInterval. The default implementation set the fetchTime to now, using the default fetchInterval.

Parameters:

url - URL of the page.

datum - datum instance to be initialized.

Returns:

adjusted page information, including all original information. NOTE: this may be a different instance than , but implementations should make sure that it contains at least all information from .

setFetchSchedule

CrawlDatum setFetchSchedule(org.apache.hadoop.io.Text url,
                            CrawlDatum datum,
                            long prevFetchTime,
                            long prevModifiedTime,
                            long fetchTime,
                            long modifiedTime,
                            int state)

Sets the fetchInterval and fetchTime on a successfully fetched page. Implementations may use supplied arguments to support different re-fetching schedules.

Parameters:

url - url of the page

datum - page description to be adjusted. NOTE: this instance, passed by reference, may be modified inside the method.

prevFetchTime - previous value of fetch time, or 0 if not available

prevModifiedTime - previous value of modifiedTime, or 0 if not available

fetchTime - the latest time, when the page was recently re-fetched. Most FetchSchedule implementations should update the value in to something greater than this value.

modifiedTime - last time the content was modified. This information comes from the protocol implementations, or is set to < 0 if not available. Most FetchSchedule implementations should update the value in to this value.

state - if STATUS_MODIFIED, then the content is considered to be "changed" before the fetchTime, if STATUS_NOTMODIFIED then the content is known to be unchanged. This information may be obtained by comparing page signatures before and after fetching. If this is set to STATUS_UNKNOWN, then it is unknown whether the page was changed; implementations are free to follow a sensible default behavior.

Returns:

adjusted page information, including all original information. NOTE: this may be a different instance than , but implementations should make sure that it contains at least all information from .

setPageGoneSchedule

CrawlDatum setPageGoneSchedule(org.apache.hadoop.io.Text url,
                               CrawlDatum datum,
                               long prevFetchTime,
                               long prevModifiedTime,
                               long fetchTime)

This method specifies how to schedule refetching of pages marked as GONE. Default implementation increases fetchInterval by 50%, and if it exceeds the maxInterval it calls forceRefetch(Text, CrawlDatum, boolean).

Parameters:

url - URL of the page

datum - datum instance to be adjusted

Returns:

adjusted page information, including all original information. NOTE: this may be a different instance than , but implementations should make sure that it contains at least all information from .

setPageRetrySchedule

CrawlDatum setPageRetrySchedule(org.apache.hadoop.io.Text url,
                                CrawlDatum datum,
                                long prevFetchTime,
                                long prevModifiedTime,
                                long fetchTime)

This method adjusts the fetch schedule if fetching needs to be re-tried due to transient errors. The default implementation sets the next fetch time 1 day in the future and increases the retry counter.

Parameters:

url - URL of the page

datum - page information

prevFetchTime - previous fetch time

prevModifiedTime - previous modified time

fetchTime - current fetch time

Returns:

adjusted page information, including all original information. NOTE: this may be a different instance than , but implementations should make sure that it contains at least all information from .

calculateLastFetchTime

long calculateLastFetchTime(CrawlDatum datum)

Calculates last fetch time of the given CrawlDatum.

Returns:

the date as a long.

shouldFetch

boolean shouldFetch(org.apache.hadoop.io.Text url,
                    CrawlDatum datum,
                    long curTime)

This method provides information whether the page is suitable for selection in the current fetchlist. NOTE: a true return value does not guarantee that the page will be fetched, it just allows it to be included in the further selection process based on scores. The default implementation checks fetchTime, if it is higher than the it returns false, and true otherwise. It will also check that fetchTime is not too remote (more than maxInterval

Parameters:

url - URL of the page

datum - datum instance

curTime - reference time (usually set to the time when the fetchlist generation process was started).

Returns:

true, if the page should be considered for inclusion in the current fetchlist, otherwise false.

forceRefetch

CrawlDatum forceRefetch(org.apache.hadoop.io.Text url,
                        CrawlDatum datum,
                        boolean asap)

This method resets fetchTime, fetchInterval, modifiedTime and page signature, so that it forces refetching.

Parameters:

url - URL of the page

datum - datum instance

asap - if true, force refetch as soon as possible - this sets the fetchTime to now. If false, force refetch whenever the next fetch time is set.

Returns:

adjusted page information, including all original information. NOTE: this may be a different instance than , but implementations should make sure that it contains at least all information from .