Removes obsolete "pipeline"

compiler/README.md

@@ -8,7 +8,7 @@ Here's a brief summary of how the compilation is done.
     The HOCON syntax used by `deepdive.conf` is interpreted by `hocon2json` and everything is converted into a single JSON config object that holds everything under the key "deepdive".
 
 2. The config object is first extended with some implied extractors, such as initializing the database and loading input tables.
-    Then, the dependencies of extractors, factors, pipelines are normalized, and their names are qualified with corresponding prefixes (by `compile-config_normalized`) to make it easier and clearer to produce the final code for execution.
+    Then, the dependencies of extractors are normalized, and their names are qualified with corresponding prefixes (by `compile-config_normalized`) to make it easier and clearer to produce the final code for execution.
     DeepDive's built-in processes for variables and factors, such as grounding, learning, inference, and calibration, are added to the config object after the normalization.
     User's original config is kept intact under "deepdive" while the normalized one is created under a different key, "deepdive_".
 

compiler/compile-check/compile-check-1-compiled_dependencies_correct

@@ -18,8 +18,6 @@ $deepdive.execution.processes | to_entries[] |
     [ .dependencies_[]?
     | select(
         (startswith("process/") and in($deepdive.execution.processes)
-      # TODO factor/ aren't really used, so remove?
-      or startswith("factor/")  and (ltrimstr("factor/") | in($deepdive.inference.factors_byName))
       or startswith("data/")    and (ltrimstr("data/")   | in($deepdive.schema.relations)
                                                            # XXX assume user is doing the right thing if schema.json is empty
                                                            or ($deepdive.schema.relations | length == 0))

compiler/compile-code/compile-code-Makefile

@@ -15,8 +15,6 @@ def mktargets: mktargets("done");
 # default commands
 define CMD_data
 endef
-define CMD_pipeline
-endef
 define CMD_process
 endef
 define CMD_factor
@@ -57,11 +55,6 @@ reset: \(keys | mktargets("reset"))
 	$(RESET) \(.key | mktarget("done"))
 "
 
-# pipelines are special
-, if .key | startswith("pipeline/") then "
-.PHONY: \(.key | mktarget)
-" else empty end
-
 )
 
 ] | join("") }

compiler/compile-code/compile-code-dataflow_dot

@@ -28,7 +28,6 @@ def nodeType: sub("/.*$"; "");
 # edge attributes by [srcType][dstType]
 (
     { ""       : { "": "color=\"#999999\"" }
-    , pipeline : { "": "style=dotted arrowhead=odiamond" }
     } * ($deepdiveDotConfig.edge_attrs // {})
 ) as $edge_attrs |
 

compiler/compile-config/compile-config-0.00-init_objects

@@ -13,8 +13,6 @@
 | .deepdive_.extraction.extractors |= . + {}
 | .deepdive_.inference             |= . + {}
 | .deepdive_.inference.factors     |= . + {}
-| .deepdive_.pipeline              |= . + {}
-| .deepdive_.pipeline.pipelines    |= . + {}
 
 # make sure our intermediate representation for execution plan set up
 | .deepdive_.execution             |= . + {}

compiler/compile-config/compile-config-1.00-qualified_names

@@ -3,7 +3,6 @@
 #
 # - extractor names are prefixed with process/*
 # - factor names are prefixed with factor/*
-# - pipeline names are prefixed with pipeline/*
 # - output_relation names are prefixed with data/* and kept under a new key output_
 # - dependencies are rewritten with qualified names under a new key dependencies_
 ##
@@ -31,16 +30,5 @@
                 )
     )
 
-# qualify names in pipelines
-| .deepdive_.pipeline.pipelines |= with_entries
-    ( .key as $p
-    | .key |= "pipeline/\(.)"
-    | .value |= map(  if $deepdive.extraction.extractors[.] then "process/\(.)"
-                    elif $deepdive.inference.factors[.]     then "factor/\(.)"
-                    else error("\(.): Neither an extractor or inference rule in pipeline \($p)")
-                    end
-                   )
-    )
-
 # turn all extractors into processes in the execution plan under compilation
 | .deepdive_.execution.processes += .deepdive_.extraction.extractors

compiler/compile-config/compile-config-2.01-grounding

@@ -702,7 +702,7 @@ def factorWeightDescriptionSqlExpr:
 
 ## from_grounding
 # A nominal process to make it easy to redo the grounding
-# TODO remove this once deepdive-do supports process groups or pipelines
+# TODO remove this once deepdive-do supports process groups
 | .deepdive_.execution.processes += {
     "process/grounding/from_grounding": {
         style: "cmd_extractor", cmd: ": no-op"

compiler/deepdive-compile

@@ -148,7 +148,7 @@ deepdive-check -a -c "$PWD"/config.json 'compiled_*' 2>&1 | sed 's/^/ /' >&2
 
 ###############################################################################
 STEP "Compiling executable code into:"
-# compile extractors and factors under process/ and factor/
+# compile extractors under process/
 pids=(--)
 for cc in "$DEEPDIVE_HOME"/util/compile-code/compile-code-*; do
     [[ -x "$cc" ]] || continue

doc/configuration.md

@@ -45,7 +45,7 @@ deepdive {
 }
 ```
 
-In this template, the global section `deepdive` contains following major sections: `db`, `schema`, `extraction`, `inference`, `calibration`. Other optional sections are `sampler` and `pipeline`.
+In this template, the global section `deepdive` contains following major sections: `db`, `schema`, `extraction`, `inference`, `calibration`. Other optional sections are `sampler` and `execution`.
 
 Links to these sections: