Running your own Hoogle on a Web Server

As promised, here is a guide on deploying Hoogle on a web server. Before doing so, you need to generate the necessary Hoogle databases, as described yesterday, and place them in the datadir configured with Cabal. Then:

• Move the hoogle binary to a location where it can act as a CGI binary, perhaps changing its name to index.cgi, if necessary. Configure the CGI program to run, possibly changing the program to be executable or adding settings somewhere.


• Copy the files from src/res in the darcs repo into a res directory located beside the binary.


• Create a file log.txt and give it global write permissions.

Now you should have Hoogle running on a web server! Some of the features, such as OpenSearch integration, won't work - but Hoogle should be usable. If anyone does get Hoogle running on a web server I'd love to hear, any feedback appreciated. In particular, if there are any tweaks required please let me know.

Hoogle Database Generation

Brief Annoucement: A new release of the Hoogle command line is out, including bug fixes and additional features. Upgrading is recommended.

Two interesting features of Hoogle 4 are working with mulitple function databases (from multiple packages), and running your own web server. Both these features aren't fully developed yet, and may change in their use, but can be used with care. This post covers how to generate your own databases, and how the web version databases are generated. Tomorrow I'm going to post on how to run your own Hoogle web server, but you'll need to generate your databases first! I'm going to walk through all the steps to create a database from the filepath library, as an example

Hoogle Databases

A Hoogle database is a set of searchable things, including text and type searching, and has a ".hoo" extension. A database may include the definitions from one package, or from multiple packages. Typically the Hoogle databases installed would include one database for each package (i.e. base.hoo, filepath.hoo), a default database (default.hoo) comprising of all the standard search items, and any number of custom databases (all.hoo) which comprise of different combinations of the other databases.

When using Hoogle, adding +name will include the given database in the search list, and -name will exclude the given package from the search. By default, Hoogle will use default.hoo, but if any +name commands are given then those databases will be used instead.

Hoogle looks for databases in the current directory, in the data directory specified by Cabal, and in any --include directories passed at the command line.

Step 1: Creating a Textbase

A Textbase is a textual representation of a function database. To generate a textbase you need to install the darcs version of Haddock, then use runhaskell Setup haddock --hoogle on your package. For filepath, this will create the file dist/doc/html/filepath/filepath.txt, which is a textbase.

Step 2: Converting a Textbase to a Database

To convert a textbase to a database use the command hoogle --convert=filepath.txt in the appropriate folder. If a package depends on any other packages, then adding +package will allow Hoogle to use the dependencies to generate a more accurate database. In the case of filepath, which depends on base, we use hoogle --convert=filepath.txt +base. This command requires base.hoo to be present.

Adding the dependencies is not strictly necessary, but will allow Hoogle to generate a more accurate database. For example, the base package defines type String = [Char], without the +base flag this type synonym would not be known to Hoogle.

We now have filepath.hoo, which can be used as a search database.

Step 3: Combining Databases

To generate a database comprising of both filepath and base, type hoogle --output=default.hoo --combine=filepath.hoo --combine=base.hoo. By combining databases you allow easy access to common groups of packages, and searching all these packages at once becomes faster than listing each database separately.

Web Version Databases

The web version uses the Hackage tarballs to generate documentation for most of its databases, but also has three custom databases:

• base - the base package is just too weird, and isn't even on hackage. A darcs version and some tweaking is required.


• keyword - the keyword database is a list of the keywords in Haskell, and is taken from the web page on the wiki.


• hackage - the hackage database is a list of all the packages on Hackage, indexed only by the package name.

All the code for generating the web version databases is found in data/generate in the Hoogle darcs repo at http://code.haskell.org/hoogle.

Future Improvements

There are two database related tasks that still need to be done: Cabal integration and indexing all of Hackage.

Bug 80: In the future I would like Hoogle databases to be generated by Cabal automatically on installing a package. Unfortunately, I don't have the time to implement such a feature currently, and even if I did implement it, I'm unlikely to ever use it. If anyone wants to work on this, please get in contact. This is mainly a project working with Cabal.

Bug 79: The other work is to index all the packages on Hackage. The problem here is generating the textbases, once they have been created the rest is fairly simple. However, to run Haddock 2 over a package requires that the package builds, and that all the dependencies are present. Unfortunatley my machine is not powerful enough to cope with the number of packages on Hackage. Hopefully at some point the machinery that builds Haddock documentation for Hackage will also generate textbases, however in the mean time if someone wants to take on the task of generating textbases for Hackage, please get in contact.

Bug Tracker

I'm not working on Hoogle full-time anymore, so am using my bug tracker to keep track of outstanding issues. In order to interact more effectively with my bug tracker, you might want to read this guide. It describes how to vote for bugs etc.

Hoogle New Features

I've now finished my Hoogle Summer of Code work, though I still intend to continue working on Hoogle when I get the chance. Before the coding period expired, I was able to add a number of new features to Hoogle. These features are all available at Hoogle, under http://haskell.org/hoogle/.

More Compact Text Searching

The old text search feature was very fast, using an on disk trie to navigate around the possible matches. The downside to this trie was the space it consumed, about half the database was devoted to it. Fortunately, I came up with an alternative way to get fast text searching (albeit slightly slower), in a lot more compact form.

Much smaller database files also mean much faster database generation, as the time spent in the IO routines is the main bottleneck.

Faster IO routines

I rewrote the underlying binary layer in Hoogle, to make it faster. It's not as fast as I would like, and I think that moving to memory-mapped files is probably a good idea. With these improvements, along with the compact text searching, I am able to generate databases in about 2 seconds (compared to about 20 seconds before).

Database Restricted Searches

Hoogle has been able to run database restricted searches for some time, but now the databases contain enough information to make it practical. By adding +package or -package to the search you can include or exclude certain packages. For example, to find out which map functions are in the containers package try map +containers. To find out which map functions are not in the containers or bytestring packages try map -containers -bytestring. I have also split out the GHC.* modules from base, so if you want to find some unboxed types in GHC's libraries try # +ghc. Note that not all the documentation links work from the GHC modules, I am still trying to fix this.

By default Hoogle searches the following packages: array, base, bytestring, cabal, containers, directory, filepath, haskell-src, hunit, keyword, mtl, parallel, parsec, pretty, process, quickcheck, random, stm, template-haskell, time, xhtml

The "ghc" package is also available if specified with +ghc and includes the GHC.* modules of base only.

Hoogle 3

I have now replaced the default Hoogle with Hoogle 4, but have copied Hoogle 3 to http://haskell.org/hoogle/3. Unfortunately, it doesn't yet work, as I need some admin help. But it will in the next few days, I hope. The only reason I can think of for using Hoogle 3 is Gtk2hs library searching, which I do want to add to Hoogle 4 when possible.

Give Me Feedback

There are quite a lot of enhancements to Hoogle that I still want to make. I have tried to list all these improvements in my bug tracker. If you find a bug, or want some feature, open an issue. If you have a particular interest in a bug, you can star it, to be informed on its progress and to indicate to me that you care.

I'm particularly interested in two pieces of feedback:

I don't use Hoogle 4 because ...

Do you use any type/name search engine? Do you want to still use Hoogle 3? Do you use Hayoo? If you use something else, what feature draws you to it? What do you dislike about Hoogle 4?

I use Hoogle 4, but my life would be nicer if ...

There are many things which effect Hoogle 4 users that I'm not aware of. If you open a bug saying what annoys you (or leave a comment and I'll do it for you) then I can keep track of this information. Even if you don't necessarily see any way to fix the problems, I'd still like to know them.

Thanks for everyone who has given feedback on Hoogle so far, it has been very useful.

GSoC Hoogle: Week 12

This week I've been trying to get Hoogle 4 to the point where it can replace Hoogle 3. This is the final official week of Google Summer of Code, but I'm planning to continue hacking Hoogle next week, and then as time allows after that.

The priority this week was getting the documentation links working. The problem was not with Hoogle - displaying the links is trivial - but ensuring that Cabal + Haddock + Hoogle + random build scripts combine to generate the correct databases. This work involved lots of little changes in lots of places, but is now working properly. Included in this work is dependency tracking of packages (so that all packages using base know that String = [Char] etc), and merging multiple databases to create a single one.

After the Hoogle database was generated correctly, I started looking at using some of the additional information present. I have now added Haddock documentation inline in the search results. If the documentation is too long to fit comfortably, Hoogle uses AJAX wizzy-ness (or more accurately, DHTML) to allow the user to expand and show all the documentation. I suspect that this will eliminate many cases of the user actually following to the Haddock webpages. This feature is fairly new, and I have pushed it out because its useful - there are still many small improvements that need to be made.

This week I also spent some time attempting to generate documentation for all the Hackage libraries. I had some success, but the computer I am currently using is years old and lacks the necessary processing power. I will tackle this at some point in the future, once I have purchased a new machine (which should be quite soon).

With all these changes, I find Hoogle 4 to be significantly more usable than Hoogle 3. Please give it a try, and give feedback. At this point I'm particularly interested in any issues that would cause you to use Hoogle 3 instead of Hoogle 4.

Hoogle 3: http://haskell.org/hoogle

Hoogle 4: http://haskell.org/hoogle/beta

If there are no major issues, I will be replacing Hoogle 4 as the standard Hoogle sometime next week.

Next week: I will be no longer doing Google Summer of Code :-) I plan to refine some of the existing bits of Hoogle, and ensure that anything I haven't done is in a bug tracker for later.

User visible changes: The web search engine now gives Haddock links and displays Haddock documentation inline.

GSoC Hoogle: Week 11

This week I've been releasing lots. Hoogle 4 is finally starting to come together, and should be a worthy replacement for Hoogle 3 very shortly. Rather than go into detail about the past week, I'm just going to give some of the bullet points:

• I have released 4 versions of the command line version of Hoogle, available on Hackage. Many bugs have been spotted by some very useful testers, and improvements have been made.


• I have released a web version of Hoogle 4, and encourage feedback.


• I have started to update the wiki Manual, which now contains some details of Hoogle's query syntax.


• I gave a talk at AngloHaskell 2008, which is available online, as slides and an audio stream. All of the other talks were excellent and are well worth listening to.


• I have started to build Hoogle documentation for all of Hackage. The machine I'm doing this on is very slow, so its not a quick process!

Next week: I'm hoping to work on generating better Hoogle databases, including a Hoogle database for the whole of Hackage. I also have a number of bugs to fix.

User visible changes: Users can download and use Hoogle, and the web interface is online.

Hoogle 4.0 web client preview

Since releasing a command line version of Hoogle 4 yesterday, I've had some useful feedback from a number of people. As a result, I have added a few bugs to the bug tracker, and fixed a few mistakes in the searching and ranking. The Hoogle on Hackage is currently 4.0.0.3 and is a recommended upgrade to all early testers.

I've now written a web interface to Hoogle 4, which has been uploaded to http://haskell.org/hoogle/beta/. This web interface is primarily so people can test searching/ranking without installing anything. There are a number of limitations:

• The links to documentation do not work - this is the most severe problem, and probably stops people permanently changing to the new version.


• The Haddock documentation is not present.


• Some database entries are duplicates.


• The Lambdabot says feature is missing.


• The Suggestion feature is incomplete.


• The AJAX style client features are not present.

The first three issues are fixed in Hoogle, but need various support through Haddock and Cabal to work. Other than these limitations, I am very interested in hearing what people think. As before, particularly regressions from Hoogle 3 or poor results/ranking.

Hoogle 4.0 release (beta, command line)

I am pleased to announce Hoogle 4.0, available on Hackage. A couple of things to note:

• This is a release of the command-line version only. It will have identical searching abilities to the web-based version, which I'm about to write.


• It currently only searches the same packages as Hoogle 3 (the final release will search more).


• It currently doesn't support the --info flag as previously described (problems with Haddock, not with Hoogle).

Walkthrough: Installation

If you have cabal-install available, it should be as simple as:

$ cabal update && cabal install hoogle

Otherwise, follow the standard Cabal/Hackage guidelines. Hoogle depends on about 4 packages on Hackage which are not available with a standard GHC install, so these will need to be built.

Walkthrough: A few searches

Here are some example searches. I have used --count=5 to limit the number of results displayed. If you are using a terminal with ANSI escape codes I recommend also passing --color to enable colored output.

$ hoogle map --count=5
Prelude map :: (a -> b) -> [a] -> [b]
Data.ByteString map :: (Word8 -> Word8) -> ByteString -> ByteString
Data.IntMap map :: (a -> b) -> IntMap a -> IntMap b
Data.IntSet map :: (Int -> Int) -> IntSet -> IntSet
Data.List map :: (a -> b) -> [a] -> [b]

$ hoogle "(a -> b) -> [a] -> [b]" --count=5
Prelude map :: (a -> b) -> [a] -> [b]
Data.List map :: (a -> b) -> [a] -> [b]
Control.Parallel.Strategies parMap :: Strategy b -> (a -> b) -> [a] -> [b]
Prelude fmap :: Functor f => (a -> b) -> f a -> f b
Control.Applicative <$> :: Functor f => (a -> b) -> f a -> f b

$ hoogle Data.Map.map --count=5
Data.Map map :: (a -> b) -> Map k a -> Map k b
Data.Map data Map k a
module Data.Map
Data.Map mapAccum :: (a -> b -> (a, c)) -> a -> Map k b -> (a, Map k c)
Data.Map mapAccumWithKey :: (a -> k -> b -> (a, c)) -> a -> Map k b -> (a, Map k c)

$ hoogle "Functor f => (a -> b) -> f a -> f b" --count=5
Prelude fmap :: Functor f => (a -> b) -> f a -> f b
Control.Applicative <$> :: Functor f => (a -> b) -> f a -> f b
Control.Monad fmap :: Functor f => (a -> b) -> f a -> f b
Control.Monad.Instances fmap :: Functor f => (a -> b) -> f a -> f b
Data.Traversable fmapDefault :: Traversable t => (a -> b) -> t a -> t b

How you can help

I've released a command line version of the search to solicit feedback. I'm interested in all comments, but especially ones of the form:

• I prefer the command line version of Hoogle 3 because ...


• When I search for ... I would expect result ... to appear, or to appear above result ...


• I was hoping for the feature ...


• It takes too long when I ...

I'm going to be accumulating Hoogle 4 bugs in my bug tracker, or by email (http://www-users.cs.york.ac.uk/~ndm/contact/) - whichever you find more convenient.

Now I'm going to start work on the Web search :-)

GSoC Hoogle: Week 10

This week I've been in Bristol, and am just about to head off to the Harbour Festival. Next week I'm heading off to AngloHaskell 2008, and will be talking about Hoogle type searching on the Saturday.

This week has been type search, yet again. There were issues with algorithmic complexity, combinatorial explosions and other fun stuff. However, its now finished. The type search is now fast enough (you can run Hoogle in Hugs against the core libraries) and gives good results. Rather than describe type searching, its easier to give an example. Searching for (a -> b) -> [a] -> [b] in Hoogle 3 gives:

Prelude.map :: (a -> b) -> [a] -> [b]
Data.List.map :: (a -> b) -> [a] -> [b]
Control.Parallel.S... parMap :: Strategy b -> (a -> b) -> [a] -> [b]
Prelude.scanr :: (a -> b -> b) -> b -> [a] -> [b]
Data.List.scanr :: (a -> b -> b) -> b -> [a] -> [b]
Prelude.scanl :: (a -> b -> a) -> a -> [b] -> [a]
Data.List.scanl :: (a -> b -> a) -> a -> [b] -> [a]
Prelude.concatMap :: (a -> [b]) -> [a] -> [b]

But in Hoogle 4 gives:

Prelude map :: (a -> b) -> [a] -> [b]
Data.List map :: (a -> b) -> [a] -> [b]
Prelude fmap :: Functor f => (a -> b) -> f a -> f b
Control.Applicative <$> :: Functor f => (a -> b) -> f a -> f b
Control.Monad fmap :: Functor f => (a -> b) -> f a -> f b
Control.Monad.Instances fmap :: Functor f => (a -> b) -> f a -> f b
Control.Applicative liftA :: Applicative f => (a -> b) -> f a -> f b
Data.Traversable fmapDefault :: Traversable t => (a -> b) -> t a -> t b
Control.Monad liftM :: Monad m => (a1 -> r) -> m a1 -> m r
Control.Parallel.Strategies parMap :: Strategy b -> (a -> b) -> [a] -> [b]

I think the new results are better. For more details, come to the AngloHaskell talk.

Next Week: I want to release a public beta of Hoogle 4 in command line form. I want to start on the web search engine and tweak the ranking algorithm. I'll also be writing up type search in the form of a presentation.

User Visible Changes: Type search works well and fast.

GSoC Hoogle: Week 9

I'm off camping for the next weekend in a couple of hours, so this is my early weekly summary. From next week, for a week and a half, I'll actually have an SSH connection so expect to see 200+ patches flow into the Hoogle repo in a few days.

This week I've been rewriting the type search. I spent 3 days writing code, type checking it, but not actually having enough written to run it. Late last night I finished the code, and this morning I debugged it. Amazingly (although actually quite commonly for Haskell) it worked with only minor tweaks. I now have a type search which should scale to large databases and provide fast and accurate searches.

All the basic tests work, and I can generate a Hoogle database for the array library. I still can't generate a Hoogle database for the base library, due to a stack overflow, but I think the cause of the stack overflow has changed and should be easy to debug.

Next Week: A public beta of the command line version is now overdue, and hopefully will happen next week. I aim to finish the actual search side of Hoogle, and move on to the web interface.

User Visible Changes: Type search works again, mostly.

GSoC Hoogle: Week 8

This week I've been travelling quite a bit, and rather busy with other things. Hopefully next week I'll be able to focus more time on Hoogle!

This week I fleshed out the final part of type search, including support for instances and alpha renaming of variables. After having implemented all the bits in the type search, I tried to convert the base libraries - and it failed, taking up too much time/memory to feasibly finish.

The type search is based around the idea of having nodes in a graph representing types, and then moving between these nodes, at a cost. In order to avoid a blow-up in the number of nodes in the graph, types are alpha-normalised and then alpha-renaming is performed afterwards. Instead of having 3 type nodes for (a,b), (c,d) and (a,a) there is just one named (a,b) and a 3 sets of alpha-renamings. All is good.

However, once you introduce instance restrictions, the types blow up. For example, from the type node a, you can move to Eq a => a, Ord a => a, Show a => a etc. The large (but feasible) number of type nodes, combined with even a small number of class names, gives a huge number of nodes. In fact, for every type variable in a node there are 2^n possible instance contexts it could take. All is bad.

Fortunately there is a solution - move instance checking outside the type graph. This makes the number of nodes feasible, and should work fairly well. It also has a few other benefits, including slightly better scoring and a simpler implementation in a few places. I also came up with a strategy for moving the cost associated with alpha-renaming into the graph search, which further simplifies things.

Of course, all this work takes time, so overall progress is slower than I would have liked. However, the results so far are promising, and the problems of scale seem to have been successfully addressed. The problem of fast and accurate type searching is hard, but hopefully Hoogle 4 will have a scalable solution that should be useful.

Next week: I want to finish the implementation of type searching, and check it works on the full base libraries. A release would be good, although may take place early in the following week.

User visible changes: Creating a database for the base library will now fail with a stack overflow. Hopefully next weeks changes will fix this!

GSoC Hoogle: Week 7

This week I've continued to improve the type searching, and generated Hoogle databases for the core libraries. I'm away from a computer all weekend until Tuesday evening, which has happened every ICFP contest for the last 3 years.

I've substantially refactored the type searching, basing it on a proper abstract Graph data type. Now the mechanisms for dealing with type search and graph traversal are separate it is much easier to express clearly what type search is doing. I've also fleshed out the type searching code so that it can accurately perform searches with all the necessary features. There are still a number of tasks to do before the type searching code is finished, but each is a fairly discrete unit of work with well-understood problems.

The other challenge for the week has been generating Hoogle databases for the core libraries - the base library and all the other libraries GHC ships with a release. With these libraries in place, it is feasible to use Hoogle to perform useful queries. The libraries are generated using a combination of Cabal, Haddock and Hoogle. I've made changes in both the Haddock and Hoogle layers so that the full base libraries can now be processed.

In order to deal with the full base libraries there are numerous GHC extensions that must be supported. In particular, Hoogle now supports multi-parameter type classes, higher-ranked types, type operators, unboxed types, unboxed tuples and NDP style arrays. All of these features are translated down into Haskell 98 types, but most closely approximate their behaviour in GHC, and can be used in searches.

Throughout the week I've been profiling the database creation code in Hoogle. The databases for the core libraries come to about 4.5Mb, and are highly optimised for performing searches - often at the cost of making them harder to create. I've halved the time to create databases during the week, using profiling to direct improvements. The time required to process the Core libraries takes 60 seconds, which is a certainly an acceptable timeframe, but could always be faster. Currently the biggest culprit in the profile is the hPutByte function:

hPutByte :: Handle -> Int -> IO ()
hPutByte hndl i = hPutChar hndl $ chr i

Since the underlying databases are written using lots of hTell and hSeek commands, it is not possible to use something like the Data.Binary library. However, if anyone has any suggestions on how to improve performance they would be gratefully received.

Next week: I want to finish off the remaining type search features, and then package up a command line release for Hackage. Hopefully Hoogle 4 will be ready for initial use by early testers.

User visible changes: Type search is more robust, but still not fully featured. Database creation if faster and robust. You can search the base libraries.

GSoC Hoogle: Week 6

This week I've been tackling type searching. I have just (in the last few minutes) got my first type search to work. At the moment type search is very limited, but all the ideas and scafolding are in place, so should now proceed relatively quickly.

In all previous versions on Hoogle, type searching was O(n), where n is the number of functions in the database. Hoogle compared the type search to each possible answer, computed a closeness score, then at the end wrote out the closest matches. This meant that before the first answer could be given, all functions had to be checked, i.e. the time for the first answer was O(n). As the Hoogle database is about to get massively bigger, this approach is insufficient.

The new version of Hoogle is much cleverer. It works by exploring a graph, following similar ideas to Dijkstra's algorithm, to reach more suitable results first. Typically, the best answers will be given without any search of the graph, and then as the graph is searched more results will appear with lower closeness. With the new scheme the complexity is O(m), where m is the number of results you want. I hope at some point after the SoC is finished to describe the algorithm properly, so others can understand it, and hopefully improve upon it.

Next week: Finishing off type searching, so it supports all the features planned. Build system work, and potentially a cabal pre-release.

User visible changes: Type search works to some degree, but not perfectly. Database debugging options (conversion and dumping to a text file) have been added.

GSoC Hoogle: Week 5

This week I was going to tackle type searching, but then realised I'm going to spend 6 hours on Friday on a train (hence the weekly update on Thursday), so can spend that time productively working on paper tackling type search. So instead of type search, I worked on a few other pieces, some of which make type search easier:

Haddock Database Generation More patches to get better output from Haddock. The code now handles class methods properly, and deals with some FFI bits.

Lazy Name Searching Searching for a name is now fairly lazy. When searching for a name, Hoogle can return the prefix of the results without doing too much computation to calculate all the results. This work is useful in its own right, but very necessary for the type searching, and can be reused.

Hoogle --info The biggest feature added this week is the --info flag. When this flag is given, Hoogle picks the first result and gives more details, including any Haddock documentation associated with the function. For example:


$ hoogle +tagsoup openurl --info
Text.HTML.Download openURL :: String -> IO String

This function opens a URL on the internet. Any http:// prefix is ignored.

> openURL "www.haskell.org/haskellwiki/Haskell"

Known Limitations:

* Only HTTP on port 80
* Outputs the HTTP Headers as well
* Does not work with all servers

It is hoped that a more reliable version of this function will be placed in a new HTTP library at some point!


Next week: Type searching! See last week for a description of what I hope to achieve.

User visible changes: The --info flag now exists.

GSoC Hoogle: Week 4

This week I've stayed in one place, and had lots of opportunity to get on with Hoogle. I've done a number of different things this week:

More on Haddock databases I fixed a number of issues with the Haddock generated Hoogle information. These patches have been submitted back to Haddock.

Binary Defer library I merged the binary defer library into the Hoogle sources, and modified it substantially. Some of the modifications were thanks to suggestions from the Haskell community, particularly David Roundy. The library is now more robust, and is being used as a solid foundation to build the rest of Hoogle on top of.

Text Searching You can now search for words, even multiple words, and the search will be performed. The text searching uses efficient data structures, scales excellently, and returns better results first.

Suggestions These improvements were detailed earlier in the week.

Next week: Type searching. I have various ideas on how to go about this, but it is the most tricky part of the whole project. I hope to come up with the perfect solution by the end of the week, but if not, will come up with something good enough for Hoogle 4 then revise it after the Summer is over (it could easily suck in a whole Summer of time if I am not careful!). Much of the low-level infrastructure is already present, so it is just the search algorithm.

User visible changes: Text searching works. A session with Hoogle as it currently stands:

> cabal haddock --hoogle
-- generates tagsoup.txt
> hoogle --convert=tagsoup.txt
Generating Hoogle database
Written tagsoup.hoo
> hoogle +tagsoup is open --color
Text.HTML.TagSoup.Type isTagOpen :: Tag -> Bool
Text.HTML.TagSoup.Type isTagOpenName :: String -> Tag -> Bool

.

Hoogle 4 New Features

I'm still developing Hoogle 4, and there are many things that don't work (such as searching for types and the web version). However, it's starting to come together, and I'm beginning to implement new features that aren't in Hoogle 3. Today I've implemented two useful features.

Multi Word Text Search

In Hoogle 3, if you entered "is just" it would be treated as a type search, exactly the same as "m a". Now, it will search for "is" and search for "just" and intersect the results. This seems to be something that people often try, so hopefully will make Hoogle more intuitive.

Intelligent Suggestions

Hoogle 3 tries to give suggestions, for example if I search for "a -> maybe a" it will helpfully suggest "a -> Maybe a". Unfortunately it's not that clever. If your search term contains a type variable (starting with a lower-case letter), which is more than one letter, it will suggest you try the capitalised version. For example, "(fst,snd) -> snd" will suggest "(Fst,Snd) -> Snd", which isn't very helpful.

The new mechanism uses knowledge about the types, arities and constructors present in the Hoogle database. Some examples:

"Just a -> a"  ===> "Maybe a -> a"
"a -> Maybe"   ===> "a -> Maybe b"
"a -> MayBe a" ===> "a -> Maybe a"
"a -> maybe a" ===> "a -> Maybe a"

.

darcs over FTP

I'm currently unable to access SSH, and suspect this situation will persist for most of the Summer. Most of my darcs repo's are behind SSH, so this presents a problem. I've been looking for a way to work with darcs over FTP, and have managed to get it going on Windows. The following are instructions for (1) me when I forget them and (2) any Windows users who want to follow the same path. If you are a Linux user, then similar information is available from this blog post.

Step 1: Install Sitecopy

Go to http://dennisbareis.com/freew32.htm and download and install SITECPY.

Add "C:\Program Files\SITECOPY" to your path.

Add "C:\Home" to a %HOME% variable.

Open up a command line and type:

c:\> mkdir home
c:\> cd home
c:\home> mkdir .sitecopy
c:\home> echo . > .sitecopyrc

Step 2: Prepare the FTP site

Go to the FTP site, and create a directory. In my particular example, I created darcs/hoogle so I could mirror the Hoogle repo.

Step 3: Configure Sitecopy

Edit the file "c:\home\.sitecopyrc" to contain:

site          hoogle
   server     ftp.york.ac.uk
   username   ndm500
   local      C:\Neil\hoogle
   remote     web/darcs/hoogle
   port       21

Obviously, substituting in your relevant details.

Step 4: Initialise Sitecopy

Type:

sitecopy --init hoogle

darcs push

Now to do a darcs push, you can type:

sitecopy --update hoogle

The first copy will take a long time, but subsequent copies should be a lot faster.

darcs pull

After all this, you can either pull using FTP, or if your FTP is also a web site, you can pull over http. For example:

darcs get http://www-users.york.ac.uk/~ndm500/darcs/hoogle/

GSoC Hoogle: Week 3

This week I've travelled a further 600 miles by train, but am now starting to get settled for the Summer, and down to work on Hoogle.

My main focus this week has been getting Haddock to generate Hoogle databases. For Haddock 0.8 I added in a --hoogle flag to generate Hoogle databases, and a similar --hoogle flag to Cabal. Unfortunately, for Haddock 2.0, the feature was removed as most of the code got rewritten. Now I've added the feature back, making extensive use of the GHC API to reduce the amount of custom pretty-printing required, and to support more Haskell features. The code has been added to the development Haddock branch, and will be present in the next release.

Most of the challenge was working with the GHC API. It's certainly a powerful body of code, but suffers from being inconsistent in various places and poorly documented. I mainly worked with the code using :i to view the API. I got bitten by various problems such as the Outputable module exporting useful functions such as mkUserStyle :: QueryQualifies -> Depth -> PprStyle, but not exporting any functions that can create a Depth value, and therefore not actually being usable. If Hoogle and Haddock could be used over the GHC API, it would substantially improve the development experience!

I've also worked more on defining the database format. I am about to start work on the implementation today. I've also added a few more command line flags, but mainly as placeholders.

Next week: Database creation and text searches (looking back I see some similarity to last week!)

User visible changes: haddock --hoogle now works.

GSoC Hoogle: Week 2

This week I submitted my PhD thesis, emptied my entire rented house of furniture, spent £96 on petrol, drove (or was driven) 400 miles, travelled a similar distance by train, have been to the north of Scotland and am currently working on a borrowed Mac in London. Needless to say, its been rather busy - but now all the excitement is over and I should be able to focus properly on Hoogle.

In the last week I've been focusing on the database, the store of all the function names and type signatures, so a very critical piece of information. I want to support fast searching, which doesn't slow down as the number of known functions increases - a nasty property of the current version. For text searching, the trie data structure has this nice property, and can deal with searching for substrings. For fuzzy type searching, things are a lot more complex. However, I think I have an algorithm which is fast (few operations), accurate (gives better matches), scalable (independent of the number of functions in the database) and lazy (returns the best results first). The idea is to have a graph of function results, and then navigate this graph to find the best match.

Most of the database work has been theoretical, but I have done some coding. In particular, I have started on the database creation code, and polished the flag argument interaction code some more. Part of the development required the Derive tool, and in doing this work I noticed a few deficiencies. In particular, if you run Windows and run derive over a UNIX line-ending file, the tool will generate a Windows line-ending file. This problem, and a few others, are now fixed.

Next week: Database creation and searching. I want text searches to work by the end of the week.

User visible changes: The --help flag prints out information on the arguments.

PS. I was looking forward to seeing some blog posts from the other Haskell summer of code students on the Haskell Planet. If any Haskell GSoC student does have a blog, please ask for it to be included!

GSoC Hoogle: Week 1

I started my Google Summer of Code project on Hoogle at the beginning of this week. In my initial application I promised to make my weekly updates via blog, so here is the first weeks report:

I've only done about half a weeks work on Hoogle this week, because I'm handing in my PhD thesis early next week, and because I'm moving house on Wednesday. I spent 14 hours on Saturday moving furniture, and many more hours than that on my thesis! I should be fully devoted to GSoC by the middle of next week.

Despite all the distractions, I did manage to start work on Hoogle. I created a new project for Hoogle at the community.haskell.org site, and an associated darcs repo at http://code.haskell.org/hoogle. I've done a number of things on Hoogle:

• Improved the developer documentation in some places


• Reorganised the repo, moving away dead files


• Work on command line flags, parsing them etc.


• Added a framework for running regression tests


• Organise the command line/CGI division

I've started work from the front, and am intending to first flesh out an API and command line client, then move on to the web front end. The biggest change from the current implementation of Hoogle will be that there is one shared binary, which will be able to function in a number of modes. These modes will include running as a web server, as a command line version, as an interactive (Hugs/GHCi style) program, documentation location etc. This will allow easier installation, and let everyone host their own web-based Hoogle without much effort.

Next week: I hope to move towards the command line client and central Hoogle database structure. I also hope to chat to the Haddock 2 people, and try and get some integration similar to Haddock 1's --hoogle flag.

User visible changes: Hoogle 4 as it currently stands is unable to run searches, although hoogle --test will run some regression tests.

Interactive TagSoup parsing

I've written quite a few programs using the tagsoup library, but have never really used the library interactively. Today I was wondering how many packages on hackage use all lower case names, compared to those starting with an initial capital. This sounds like a great opportunity to experiment! The rest of this post is a GHCi transcript, with my comments on what I'm doing prefixed with -- characters.

$ ghci
GHCi, version 6.8.2: http://www.haskell.org/ghc/  :? for help
Loading package base ... linking ... done.
-- load some useful packages
Prelude> :m Text.HTML.TagSoup Text.HTML.Download Data.List Data.Char Data.Maybe
Prelude Data.Maybe Data.Char Data.List Text.HTML.Download Text.HTML.TagSoup>
-- ouch, that prompt is a bit long - we can use :set prompt to shorten it
-- side note: I actually supplied the patch for set prompt :)
   :set prompt "Meep> "
-- lets download the list of packages
Meep> src <- openURL "http://hackage.haskell.org/packages/archive/pkg-list.html"
... src scrolls pass the screen ...
-- parse the file, dropping everything before the packages
Meep> let parsed = dropWhile (~/= "<h3>") $ parseTags src
-- grab the list of packages
Meep> let packages = sort [x | a:TagText x:_ <- tails parsed, a ~== "<a href>"]
-- now we can query the list of packages
Meep> length packages
648
Meep> length $ filter (all isLower) packages
320
Meep> length $ filter ('_' `elem`) packages
0
Meep> length $ filter ('-' `elem`) packages
165
Meep> length $ filter (any isUpper . dropWhile isUpper) packages
100
Meep> length $ filter (isPrefixOf "hs" . map toLower) packages
47
Meep> length $ filter (any isDigit) packages
37
Meep> reverse $ sort $ map (\(x:xs) -> (1 + length xs,x)) $ group $ sort $ conca
t packages
[(484,'e'),(374,'a'),(346,'r'),(336,'s'),(335,'t'),(306,'i'),(272,'l'),(248,'c')
,(247,'n'),(240,'o'),(227,'p'),(209,'h'),(185,'-'),(171,'m'),(159,'d'),(126,'g')
,(112,'b'),(96,'u'),(87,'y'),(78,'k'),(76,'f'),(74,'x'),(58,'S'),(53,'H'),(35,'w
'),(33,'v'),(29,'q'),(29,'L'),(27,'A'),(26,'F'),(24,'D'),(23,'C'),(22,'T'),(16,'
P'),(16,'M'),(16,'I'),(16,'G'),(13,'B'),(12,'W'),(12,'3'),(12,'2'),(10,'O'),(9,'
R'),(9,'1'),(8,'z'),(8,'j'),(8,'E'),(7,'X'),(7,'U'),(7,'N'),(6,'Y'),(6,'V'),(5,'
J'),(4,'Q'),(4,'5'),(4,'4'),(3,'Z'),(3,'8'),(3,'6'),(1,'9')]

We can see that loads of packages use lowercase, lots of packages use upper case, quite a few use CamelCase, quite a few start with "hs", none use "_", but lots use "-". The final query figures out which is the most common letter in hackage packages, and rather unsurprisingly, it roughly follows the frequency of English letters.

TagSoup and GHCi make a potent combination for obtaining and playing with webpages.