search

YannBrrd / elasticsearch-entity-resolution

Elasticsearch entity resolution plugin based on Duke
http://yannbrrd.github.io
Apache License 2.0
210 stars 52 forks source link

Custom comparator #14

Open iDmple opened 9 years ago

iDmple commented 9 years ago

Hi Yann,

I need to use WeightedLevenshtein which is available in Duke.

In Duke I set the parameters like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein$DefaultWeightEstimator"
        name="weight">
    <param name="digit-weight" value="3.8"/>
</object>

then create the comparator named wl like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein"
        name="wl">
    <param name="estimator" value="weight"/>
</object>

and then it is available to be used:

    <property>
        <name>StreetAndNr</name>
        <comparator>wl</comparator>
        <low>0.3</low>
        <high>0.55</high>
    </property>

I could not find how to do this in your documentation.

Could you please provide an example?

Many thanks!

YannBrrd commented 9 years ago

Hi,

Did you succeed finally ?

Is it linked to #15 ?

Le mer. 23 sept. 2015 09:53, iDmple notifications@github.com a écrit :

Hi Yann,

I need to use WeightedLevenshtein which is available in Duke.

In Duke I set the parameters like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein$DefaultWeightEstimator" name="weight">

then create the comparator named wl like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein" name="wl">

and then it is available to be used:

<property>
    <name>StreetAndNr</name>
    <comparator>wl</comparator>
    <low>0.3</low>
    <high>0.55</high>
</property>

I could not find how to do this in your documentation.

Could you please provide an example?

Many thanks!

— Reply to this email directly or view it on GitHub https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14.

Cordialement, Yann Barraud

iDmple commented 9 years ago

No, I didn't. I'm not sure what syntax to use because the weights are set in a subclass of the comparator, unlike the Geo comparator you used in your examples. It isn't linked to #15.

On Sep 24, 2015, 08:59, at 08:59, Yann Barraud notifications@github.com wrote:

Hi,

Did you succeed finally ?

Is it linked to #15 ?

Le mer. 23 sept. 2015 09:53, iDmple notifications@github.com a écrit :

Hi Yann,

I need to use WeightedLevenshtein which is available in Duke.

In Duke I set the parameters like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein$DefaultWeightEstimator" name="weight">

then create the comparator named wl like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein" name="wl">

and then it is available to be used:

<property>
    <name>StreetAndNr</name>
    <comparator>wl</comparator>
    <low>0.3</low>
    <high>0.55</high>
</property>

I could not find how to do this in your documentation.

Could you please provide an example?

Many thanks!

— Reply to this email directly or view it on GitHub

https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14.

Cordialement, Yann Barraud

Reply to this email directly or view it on GitHub: https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-142834035

YannBrrd commented 9 years ago

ok

Le jeu. 24 sept. 2015 à 09:08, iDmple notifications@github.com a écrit :

No, I didn't. I'm not sure what syntax to use because the weights are set in a subclass of the comparator, unlike the Geo comparator you used in your examples. It isn't linked to #15.

On Sep 24, 2015, 08:59, at 08:59, Yann Barraud notifications@github.com wrote:

Hi,

Did you succeed finally ?

Is it linked to #15 ?

Le mer. 23 sept. 2015 09:53, iDmple notifications@github.com a écrit :

Hi Yann,

I need to use WeightedLevenshtein which is available in Duke.

In Duke I set the parameters like this:

<object

class="no.priv.garshol.duke.comparators.WeightedLevenshtein$DefaultWeightEstimator"

name="weight">

then create the comparator named wl like this:

<object class="no.priv.garshol.duke.comparators.WeightedLevenshtein" name="wl">

and then it is available to be used:

StreetAndNr wl 0.3 0.55

I could not find how to do this in your documentation.

Could you please provide an example?

Many thanks!

— Reply to this email directly or view it on GitHub

https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14.

Cordialement, Yann Barraud

Reply to this email directly or view it on GitHub:

https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-142834035

— Reply to this email directly or view it on GitHub https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-142835811 .

Cordialement, Yann Barraud

NTBTI commented 8 years ago

Hi Yann I just came across this while searching for ways to use WeightedLevenshtein and am having the same issue as OP. I'm using the most recent version....is there a way to use this with your excellent tool?

iDmple commented 8 years ago

Hi, I still need this functionality.

I also made a lot of bug fixes to the original code. I need to find time for a pull request. Tell me if it's urgent.

Best regards, Nathalie

On Apr 2, 2016, 05:26, at 05:26, NTBTI notifications@github.com wrote:

Hi Yann I just came across this while searching for ways to use WeightedLevenshtein and am having the same issue as OP. I'm using the most recent version....is there a way to use this with your excellent tool?

You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-204638129

ghost commented 8 years ago

Sorry iDmple....are you saying you have WeightedLevenshtein working and just looking to do a PR? If you have it committed I would love to try it out and see how you implemented it. If not I hope Yann can give some guidance as we'd really like to implement this feature for an upcoming release.

iDmple commented 8 years ago

No, I don't. However I have many other bug fixes. I should really review everything and PR them.

Cheers

On Apr 8, 2016, 00:51, at 00:51, AbsoluteResults notifications@github.com wrote:

Sorry iDmple....are you saying you have WeightedLevenshtein working and just looking to do a PR? If you have it committed I would love to try it out and see how you implemented it. If not I hope Yann can give some guidance as we'd really like to implement this feature for an upcoming release.

You are receiving this because you authored the thread. Reply to this email directly or view it on GitHub: https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-207126604

ghost commented 8 years ago

I forked the project and made very simple changes so WeightedLevenshtein can be called using all the Duke options (digit-weight, punctuation-weight, etc). Should be accessible through my profile. If a pull request is desired I am more than happy to do that.

YannBrrd commented 8 years ago

Please feel free to do so !

Le mar. 26 avr. 2016 19:40, AbsoluteResults notifications@github.com a écrit :

I forked the project and made very simple changes so WeightedLevenshtein can be called using all the Duke options (digit-weight, punctuation-weight, etc). Should be accessible through my profile. If a pull request is desired I am more than happy to do that.

— You are receiving this because you commented.

Reply to this email directly or view it on GitHub https://github.com/YannBrrd/elasticsearch-entity-resolution/issues/14#issuecomment-214823526

Cordialement, Yann Barraud

ghost commented 8 years ago

Done and done! #28 thanks so much